final-project/src/q1.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "gpuClass": "standard",
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ThPDzzjGIcBX",
        "outputId": "db12740f-f968-4860-f0b2-dc9fccb70421"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1.13.1+cu116\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.9/1.9 MB\u001b[0m \u001b[31m1.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m9.4/9.4 MB\u001b[0m \u001b[31m53.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m4.5/4.5 MB\u001b[0m \u001b[31m90.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m3.2/3.2 MB\u001b[0m \u001b[31m83.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m873.9/873.9 KB\u001b[0m \u001b[31m63.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m565.0/565.0 KB\u001b[0m \u001b[31m8.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m280.2/280.2 KB\u001b[0m \u001b[31m23.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Building wheel for torch-geometric (setup.py) ... \u001b[?25l\u001b[?25hdone\n"
          ]
        }
      ],
      "source": [
        "import os\n",
        "import torch\n",
        "\n",
        "os.environ['TORCH'] = torch.__version__\n",
        "os.environ['CUDA'] = torch.version.cuda\n",
        "print(os.environ['TORCH'])\n",
        "\n",
        "!pip install -q pyg-lib torch-scatter torch-sparse torch-cluster torch-spline-conv -f https://data.pyg.org/whl/torch-${TORCH}.html\n",
        "!pip install -q torch-geometric\n",
        "\n",
        "from torch.nn import ReLU, Softmax\n",
        "import torch.nn.functional as F\n",
        "import torch_geometric\n",
        "from torch_geometric.datasets import DBLP\n",
        "import torch_geometric.transforms as T\n",
        "from torch_geometric.nn import Sequential, Linear, GCNConv, SAGEConv, GATConv, GATv2Conv, to_hetero\n",
        "from torch_geometric.utils import softmax\n",
        "\n",
        "import math\n",
        "import time\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "\n",
        "import warnings\n",
        "warnings.filterwarnings(\"ignore\")\n",
        "\n",
        "sns.reset_defaults()\n",
        "sns.set_theme(rc={\"figure.dpi\": 72, \"savefig.dpi\": 300,\n",
        "              \"figure.autolayout\": True})\n",
        "sns.set_style(\"ticks\")\n",
        "sns.set_context(\"paper\")\n",
        "\n",
        "np.set_printoptions(precision=4)\n",
        "\n",
        "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "dataset = DBLP(\"q1\", transform=T.Constant(node_types='conference'))\n",
        "data = dataset[0]"
      ],
      "metadata": {
        "id": "xNKZJBR9LROg",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "138a58ed-2a5f-45c1-bd68-ab43c47a9888"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "Downloading https://www.dropbox.com/s/yh4grpeks87ugr2/DBLP_processed.zip?dl=1\n",
            "Extracting q1/raw/DBLP_processed.zip\n",
            "Processing...\n",
            "Done!\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Part a\n",
        "num_classes = len(np.unique(data[\"author\"].y))\n",
        "print(dataset)\n",
        "print(\"Number of graphs:\", len(dataset))\n",
        "print(\"Number of nodes:\", data.num_nodes)\n",
        "print(\"Number of edges:\", data.num_edges)\n",
        "print(\"Number of node features:\", data.num_node_features)\n",
        "print(\"Number of edge features:\", data.num_edge_features)\n",
        "print(\"Number of classes:\", num_classes)\n",
        "print(\"Average node degree:\", round(data.num_edges / data.num_nodes, 2))\n",
        "print(\"Number of training nodes:\", data[\"author\"].train_mask.sum().item())\n",
        "print(\"Number of validation nodes:\", data[\"author\"].val_mask.sum().item())\n",
        "print(\"Number of test nodes:\", data[\"author\"].test_mask.sum().item())\n",
        "print(\"Is directed:\", data.is_directed())\n",
        "print(\"Has isolated nodes:\", data.has_isolated_nodes())\n",
        "print(\"Has self loops:\", data.has_self_loops())\n",
        "print()\n",
        "print(\"Other information:\")\n",
        "print(data)\n",
        "data = data.to(device)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uPsS2-g_LtM3",
        "outputId": "a4986ced-091b-437e-c768-2ba78fa00372"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "DBLP()\n",
            "Number of graphs: 1\n",
            "Number of nodes: 26128\n",
            "Number of edges: 239566\n",
            "Number of node features: {'author': 334, 'paper': 4231, 'term': 50, 'conference': 1}\n",
            "Number of edge features: {('author', 'to', 'paper'): 0, ('paper', 'to', 'author'): 0, ('paper', 'to', 'term'): 0, ('paper', 'to', 'conference'): 0, ('term', 'to', 'paper'): 0, ('conference', 'to', 'paper'): 0}\n",
            "Number of classes: 4\n",
            "Average node degree: 9.17\n",
            "Number of training nodes: 400\n",
            "Number of validation nodes: 400\n",
            "Number of test nodes: 3257\n",
            "Is directed: False\n",
            "Has isolated nodes: False\n",
            "Has self loops: False\n",
            "\n",
            "Other information:\n",
            "HeteroData(\n",
            "  \u001b[1mauthor\u001b[0m={\n",
            "    x=[4057, 334],\n",
            "    y=[4057],\n",
            "    train_mask=[4057],\n",
            "    val_mask=[4057],\n",
            "    test_mask=[4057]\n",
            "  },\n",
            "  \u001b[1mpaper\u001b[0m={ x=[14328, 4231] },\n",
            "  \u001b[1mterm\u001b[0m={ x=[7723, 50] },\n",
            "  \u001b[1mconference\u001b[0m={\n",
            "    num_nodes=20,\n",
            "    x=[20, 1]\n",
            "  },\n",
            "  \u001b[1m(author, to, paper)\u001b[0m={ edge_index=[2, 19645] },\n",
            "  \u001b[1m(paper, to, author)\u001b[0m={ edge_index=[2, 19645] },\n",
            "  \u001b[1m(paper, to, term)\u001b[0m={ edge_index=[2, 85810] },\n",
            "  \u001b[1m(paper, to, conference)\u001b[0m={ edge_index=[2, 14328] },\n",
            "  \u001b[1m(term, to, paper)\u001b[0m={ edge_index=[2, 85810] },\n",
            "  \u001b[1m(conference, to, paper)\u001b[0m={ edge_index=[2, 14328] }\n",
            ")\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def train(model, optimizer):\n",
        "    model.train()\n",
        "    optimizer.zero_grad()\n",
        "    out = model(data.x_dict, data.edge_index_dict)\n",
        "    mask = data[\"author\"].train_mask\n",
        "    loss = F.cross_entropy(out[\"author\"][mask], data[\"author\"].y[mask])\n",
        "    loss.backward()\n",
        "    optimizer.step()\n",
        "    return float(loss)\n",
        "\n",
        "\n",
        "def test(model):\n",
        "    model.eval()\n",
        "    pred = model(data.x_dict, data.edge_index_dict)[\"author\"].argmax(dim=-1)\n",
        "    accs = []\n",
        "    for split in ['train_mask', 'val_mask', 'test_mask']:\n",
        "        mask = data[\"author\"][split]\n",
        "        acc = (pred[mask] == data[\"author\"].y[mask]).sum() / mask.sum()\n",
        "        accs.append(float(acc))\n",
        "    return accs\n",
        "\n",
        "\n",
        "def epoch_plot(model, optimizer, max_epoch=100, save=\"epoch.png\"):\n",
        "    values = []\n",
        "    for i in range(max_epoch):\n",
        "        loss = train(model, optimizer)\n",
        "        accs = test(model)\n",
        "        values.append([i+1, loss, *accs])\n",
        "    df = pd.DataFrame(values, columns=[\"Epoch\", \"Loss\", \"Train\", \"Validation\", \"Test\"])\n",
        "    fig, ax = plt.subplots(1, 2, figsize=(10, 5))\n",
        "    sns.lineplot(data=df, x=\"Epoch\", y=\"Loss\", ax=ax[0])\n",
        "    sns.lineplot(data=df, x=\"Epoch\", y=\"Train\", label=\"Train\", ax=ax[1])\n",
        "    sns.lineplot(data=df, x=\"Epoch\", y=\"Validation\", label=\"Validation\", ax=ax[1])\n",
        "    sns.lineplot(data=df, x=\"Epoch\", y=\"Test\", label=\"Test\", ax=ax[1])\n",
        "    ax[1].set(ylabel=\"Accuracy\")\n",
        "    fig.savefig(save)\n",
        "    plt.show()\n",
        "\n",
        "\n",
        "def eval(model, optimizer, max_epoch=100):\n",
        "    values = []\n",
        "    t = time.time()\n",
        "    for i in range(max_epoch):\n",
        "        loss = train(model, optimizer)\n",
        "        accs = test(model)\n",
        "        values.append([i+1, loss, *accs])\n",
        "    t = time.time() - t\n",
        "    df = pd.DataFrame(values, columns=[\"Epoch\", \"Loss\", \"Train\", \"Validation\", \"Test\"])\n",
        "    acc = df.iloc[-1, -1]\n",
        "    print(f\"Accuracy: {acc:.4f}\")\n",
        "    print(f\"Computation Time: {t:.4f}s\")"
      ],
      "metadata": {
        "id": "ppTb9miV3SWz"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Part b, GCN\n",
        "\n",
        "def gcn(hidden_channels, num_gcn=2, num_lin=1):\n",
        "    torch.manual_seed(1234567)\n",
        "    layers = []\n",
        "    for _ in range(num_gcn):\n",
        "        layers.append((SAGEConv((-1, -1), hidden_channels, aggr=\"mean\", normalize=True), 'x, edge_index -> x'))\n",
        "        layers.append(ReLU(inplace=True))\n",
        "    for _ in range(num_lin):\n",
        "        layers.append((Linear(-1, num_classes), 'x -> x'))\n",
        "    model = Sequential('x, edge_index', layers)\n",
        "    model = to_hetero(model, data.metadata()).to(device)\n",
        "    return model\n"
      ],
      "metadata": {
        "id": "QZN7vtMrUG0H"
      },
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model = gcn(hidden_channels=64, num_gcn=2, num_lin=1)\n",
        "optimizer = torch.optim.Adam(model.parameters(), lr=0.005)\n",
        "epoch_plot(model, optimizer, max_epoch=250, save=\"gcn_epoch.png\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 370
        },
        "id": "y2K8xxc-_8CG",
        "outputId": "6a05a0ec-c24b-480c-88aa-a5838a87012d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x360 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def gcn_hidden_plot(hiddens, max_epoch=100, save=\"gcn_hidden.png\"):\n",
        "    data = []\n",
        "    for hid in hiddens:\n",
        "        model = gcn(hidden_channels=hid, num_gcn=2, num_lin=1)\n",
        "        optimizer = torch.optim.Adam(model.parameters(), lr=0.005)\n",
        "        values = []\n",
        "        for epo in range(max_epoch):\n",
        "            loss = train(model, optimizer)\n",
        "            accs = test(model)\n",
        "            values.append([epo+1, loss, *accs])\n",
        "        data.append([hid, *values[-1][1:]])\n",
        "    df = pd.DataFrame(data, columns=[\"Hidden Channels\", \"Loss\", \"Train\", \"Validation\", \"Test\"])\n",
        "    fig, ax = plt.subplots(1, 2, figsize=(10, 5))\n",
        "    sns.lineplot(data=df, x=\"Hidden Channels\", y=\"Loss\", marker=\"o\", ax=ax[0])\n",
        "    sns.lineplot(data=df, x=\"Hidden Channels\", y=\"Train\", label=\"Train\", marker=\"o\", ax=ax[1])\n",
        "    sns.lineplot(data=df, x=\"Hidden Channels\", y=\"Validation\", label=\"Validation\", marker=\"o\", ax=ax[1])\n",
        "    sns.lineplot(data=df, x=\"Hidden Channels\", y=\"Test\", label=\"Test\", marker=\"o\", ax=ax[1])\n",
        "    ax[0].set(xscale=\"log\")\n",
        "    ax[1].set(ylabel=\"Accuracy\", xscale=\"log\")\n",
        "    fig.savefig(save)\n",
        "    plt.show()\n",
        "    return df.round(4).head(None)\n",
        "\n",
        "\n",
        "gcn_hidden_plot([2, 4, 8, 16, 32, 64, 128, 256, 512, 1024])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 716
        },
        "id": "vy1pyMDUEyVs",
        "outputId": "383b2e0a-4b44-4d52-cb9c-02094b14fd85"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x360 with 2 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAskAAAFhCAYAAACCkjfzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzde3TT9f0/8OcnSZP0lqZt2rSlpVfojbaUQlG8IMrwrniZUhFkKu7rZtUN0R/O4TZF3QQdwtevuqmboihs4GU61M37jTsU2tJLWmjTlqa3pNckzeX3R0tLAbGXJJ+meT7O6UnTtP289Gjy7Duv9+stOJ1OJ4iIiIiIaIBE7AKIiIiIiMYbhmQiIiIiolMwJBMRERERnYIhmYiIiIjoFAzJRERERESnYEgmIiIiIjqFTOwCRiI1NVXsEoiIXKqsrEzsEkaMz8VENJH80POwV4VkwDtfUIiIzsSbwyafi4loIjjb8zDbLYiIiIiITsGQTERERER0CoZkIiIiIqJTMCQTEREREZ2CIZmIiIiI6BQMyUREREREp2BIJiIiIiI6BUMyEREREdEpGJKJiIiIiE4xoUOy2WLD0YZ2vP91FY42tMNssYldEhGRR/X29mLRokWYOXMmduzYcdrjn332GW6++WYsWrQIRUVFIlToXmK+Doj9GuSr/+y8Nq/tKl53LPVIHG/txn3rPoPDCUgEYP2KeUiIVoldFhGRx8hkMjz33HN4++23T3vMbrdj/fr12LRpE7q6unD//fdj8+bNIlTpPqe+Djxz/1w4AdjsDrdeVyaVQADw6z9/4fFri319XpvXFvParsx6EzokH9I1w+Hs+9zhBA7rmhmSicinCIKAyMjIMz529OhRJCQkICgoCEFBQbDZbLBYLFAoFAPfs2HDBmzcuNFT5brcqa8Du0sbcaiyGcVVLW697rSkcExL0YhybbGvz2vz2mJe25VZb0KH5KxkDSQCBv66yErWiF0SEdG4YTKZoFINvpioVCoYjUZotdqBrxUWFqKwsHDIz6WmpnqsxrE69XXg3GnRWPQTz9R/tKEdmz86Isq1xb4+r81ri3VtV2a9CR2So8ICsH7FPOwpPY7MJA20YQFil0RENG6EhISgo6Nj4H5HRwfUarWIFbleVFgAnvjF+Sg71oq8NK1HXwdOvAYd1jUjK9nzr0FiXp/X5rUnwrUndEhWKmRIiFahvdOCF/5ZhGd/NVfskoiIxo34+HgcPXoU3d3d6OrqglQqHdJqMREoFTK8/N5hLLsyA/Eebrc78RokVpufmNfntXntiXDtCR2ST0iaFIKjx9uhqzNiSlyo2OUQEXnUfffdh8OHDyMgIABFRUVQq9WYP38+kpKScM899+BnP/sZBEHAqlWrxC7V5do6zKioNSIilO8kEtHI+ERIDgqQIz0hDHtKDQzJRORz1q9f/4OPXXLJJbjkkks8WI1n6fQmBCpliApnSCaikZnQc5JPlpcWib2ljWKXQUREHlSpNyI5Vg1BEMQuhYi8jM+E5JnpWpTXtsHUaRG7FCIi8pDK2r6QTEQ0Uj4TkhOiVQhTKbGvzCB2KURE5CG6OhNSYkPELoOIvJDPhGRBEDAzXYs9bLkgIvIJpk4Lmo09SOFKMhGNgs+EZADIS9Ni3xED7CeOZiEiogmrUm9EgFKGqPBAsUshIi/kUyE5Z4oGZqsN5cfaxC6FiIjcrFJvRNKkEEgk3LRHRCPnUyE5QOmHzKRw7DnClgsioolOpzex1YKIRs2nQjIA9iUTEfkInZ6TLYho9HwuJOelaVFVZ0KLqUfsUoiIyE3au6wwtPVwsgURjZrPheTYyCBowwKw7whHwRERTVSVeiP8FVLEaILELoWIvJTPheSBUXDsSyYimrB0eiOSJqm5aY+IRs3nQjLQ15e8v6wJNrtD7FKIiMgNdHoTktlqQURj4JMheVpyOOx2B0qrW8UuhYiI3KBSb+RkCyIaE58MyUq5DFkpGk65ICKagDq6rWhs7WZIJqIx8cmQDIB9yUREE5ROb4RSLkVMBDftEdHo+XRIrjneAUNrt9ilEBGRC1XqTUiMCYGUm/aIaAx8NiRHhQdiUkQQ9nI1mYhoQtHpjUiJY6sFEY2Nz4Zk4MTpe5yXTEQ0kfQdR83JFkQ0Nj4ekiNxsLIJ1l672KUQEZELdPb0oqGli8dRE9GY+XRIzkwKhwDgcFWL2KUQEZEL6PRGyP2kiI0MFrsUIvJyPh2S/WRS5EyJwF6OgiMimhB0eiOSYlTctEdEY+bTIRk40ZfMkExENBH09SOz1YKIxs7nQ3Jemhb1zV2ob+oUuxQiIhqjSr2R/chE5BI+H5IjQv2REK3iwSJERF6uq6cX9c1dHP9GRC7h8yEZAPLSIrGXo+CIiLxaVZ0JcpkEcZE8aY+Ixo4hGX19yYd0zTBbbGKXQkREo1SpN/adtCflSxsRjZ1bn0m2bNmCRYsWYcmSJaitrR3y2Pbt23HDDTfgpz/9Kf7+97+7s4wflZYQBrlMgiJds6h1EBHR6On0JiTzEBEichG3hWSj0YitW7di06ZNWLlyJdauXTvk8RdffBGvv/463n77bbz99tuwWq3uKuVHyaQSTE+N5JQLIiIvVqk3crIFEbmM20JyUVER8vPzIZPJkJ2djerq6iGPJyUlobu7G2azGUqlElKp1F2lDMvMNC32ljbC6XSKWgcREY1ct7kX9c2d3LRHRC4jc9cvNplMCAkZfNvr1PB56aWXYuHChZBKpbjjjjtOC8kbNmzAxo0b3VXeafLSIrH+7f2obezA5CiVx65LRERjV1VngkwqQZyWJ+0RkWu4bSVZpVKhvb198EKSwUt1dnbixRdfxI4dO/DJJ5/g448/Rn19/ZCfLywsRFlZ2ZAPdwpVKZESG4I9nHJBROR1KvUmJESrIOOmPSJyEbc9m+Tk5GD37t2w2+0oLi5GfHz84EUlEvj5+SEgIAByuRxKpRKdneIf5pGXrsVezksmIvI6ujr2IxORa7mt3UKtVmPhwoVYvHgxZDIZ1qxZg23btiE2Nhb5+fm47rrrcPPNN0MQBEyfPh1Tp051VynDNjNdi3/8twLd5l4EKP3ELoeIiIZJpzfi2gtTxC6DiCYQt4VkACgoKEBBQcHA/ZNXk5ctW4Zly5a58/IjNiUuFAFKPxwob8Kc7BixyyEiomHosdigN3QihePfiMiF2Lx1EqlEQF4aR8EREXmTqjoTpBIJN10TkUsxJJ9iZn9fMkfBERF5B53eiIToYPjJ+JJGRK7DZ5RT5KZGwthhQXV9+49/MxERiU5XZ0IyN+0RkYsxJJ9CFShHanwYWy6IiLwET9ojIndgSD6DvHT2JRMReQOzxQZ9YwdDMhG5HEPyGcxM06LsWCs6uq1il0JERGdRXd8OiURAfDRP2iMi12JIPoOkSSFQByuwv4yn7xERjWeVeiMmR6ngJ5OKXQoRTTAMyWcgCALy0rRsuSAiGud40h4RuQtD8g/oO6LaAIeDo+CIiMYrnd7EQ0SIyC0Ykn/A9CkR6OrpRaXeKHYpRER0BpZeO2oaOzj+jYjcgiH5BwT6+yEjMZwtF0RE41R1vQkCgIRonrRHRK7HkHwWMzkKjoho3NLpTZgcFQy5HzftEZHrMSSfRV66FhW1RrR1mMUuhYiITqHjISJE5EYMyWcxWRuMiFB/joIjIhqHKvVG9iMTkdswJJ+FIAiYmabFnlKGZCKi8cTaa0fN8Q4kc7IFEbkJQ/KPmJmuxb4yA+x2h9ilEBFRv6MN7XACSIxhSCYi92BI/hHZKRpYrHYcOdYmdilERNRPpzdisjYYCm7aIyI3YUj+EUqFDFnJHAVHRN5ry5YtWLRoEZYsWYLa2tohj73xxhu4+eabceutt6K4uFikCkeuUm9iqwURuRVD8jDMTOcR1UTknYxGI7Zu3YpNmzZh5cqVWLt27cBjzc3NeP/99/Hmm2/iz3/+M5566ikRKx2ZSk62ICI3Y0gehpnpWhxtaEezsUfsUoiIRqSoqAj5+fmQyWTIzs5GdXX1wGN1dXVISUmBVCqFRqNBY2MjrFariNUOT6/Njprj7UiexJBMRO7DkDwMMRFBiNYEYu8RriYTkXcxmUwICRlsS3A6nQOfT548GcXFxejp6UF1dTXq6+thMpmG/PyGDRuQmpo65ENsRxva4XA4kTiJJ+0RkfvIxC7AW5xoubj0nASxSyEiGjaVSoWysrKB+xLJ4NpIaGgoli9fjuXLlyMmJgZTp05FaGjokJ8vLCxEYWHhkK+JHZR1ehNitcFQyvkSRkTuw5XkYZqZpsXBiib02uxil0JENGw5OTnYvXs37HY7iouLER8fP+TxK664Aps2bcI999yDpKQkyGTjP3iyH5mIPGH8PxuOE9OSw2F3ACVVrciZGiF2OUREw6JWq7Fw4UIsXrwYMpkMa9aswbZt2xAbG4v8/HysWLECTU1NCAkJwerVq8Uud1h0eiPmzYwTuwwimuAYkodJ7idFzhQN9hxpZEgmIq9SUFCAgoKCgfsnryavW7dOjJJGrdfmwNGGDm7aIyK3Y7vFCHAUHBGRuI4db4fd4UDSJM5IJiL3Ykgegbw0LfSGThxv6RK7FCIin6TTmxAbGQR/Bd8IJSL3YkgeAW1YAOK0wdjL1WQiIlHo9EYkc9MeEXkAQ/IIzUzXYs8Rg9hlEBH5JE62ICJPYUgeoZnpkSiqaIKll6PgiIg8yWZ34GhDO5LZj0xEHsCQPELpCeGQSiU4VNksdilERD6l5ngHbHZu2iMiz2BIHiE/mQTTp0awL5mIyMN0eiNiNEEIUPqJXQoR+QCG5FHo60tuhNPpFLsUIiKfwX5kIvIkhuRRyEuLxPGWbtQ1dYpdChGRz9DpTUiJY6sFEXkGQ/IohIf4IykmBHtKOeWCiMgT7HYHqutNPGmPiDyGIXmU8tIj2ZdMROQhNY0dsNq4aY+IPIcheZRmpmtxuKoZPRab2KUQEU14Or0JMZpABPpz0x4ReQZD8iilTg6FUi7DwYomsUshIprwdNy0R0QexpA8SlKpBDNSI7GHLRdERG5XqTciOZatFkTkOQzJY5CXrsXeUo6CIyJyJ7vdgar6diRzJZmIPIgheQxmpEaipd2MY8c7xC6FiGjC0jd1wtprZ0gmIo9iSB4DdbACU+LUbLkgInIjnd6I6PBABHHTHhF5EEPyGM1M0zIkExG5UaXexH5kIvI4huQxykvXovRoKzp7esUuhYhoQqqsNbLVgog8jiF5jFJi1QgO8MOBcp6+R0TkanaHE1X1JqRwJZmIPIwheYwkEgF5bLkgInKL+qZOWKzctEdEnseQ7AIz07TYe8QAh4Oj4IiIXKlSb4Q2LADBAXKxSyEiH8OQ7AK5qRFo77Sgqs4kdilERBNKJU/aIyKRMCS7QFCAHGkJYdhzhC0XRESupONkCyISCUOyi8xMZ18yEZErORxOVNVxsgURiYMh2UVmpmtRXtMGU6dF7FKIiCaE+uZO9FjsbLcgIlEwJLtIQrQKocFK7C/jKDgiIleo1JsQGeoPVSA37RGR5zEku4ggCP0tFwzJRESuoNOz1YKIxOPWkLxlyxYsWrQIS5YsQW1t7ZDHWlpacO+992Lp0qVYsWKFO8vwmJnpkdhX1gg7R8EREY1Zpd7ITXtEJBqZu36x0WjE1q1bsXnzZpSUlGDt2rVYv379wONPPfUUHnjgAUyePNldJXhczpQI9FhsqKhpQ1pCmNjlEBF5LYfDCZ3ehBsvniJ2KUTko9y2klxUVIT8/HzIZDJkZ2ejurp64DG73Y6qqips2LABt956Kz788EN3leFRAUo/ZCSGc8oFEdEYHW/pQo/Fxk17RCQat4Vkk8mEkJDBt8mczsEWhJaWFpSVleHuu+/GSy+9hJdeeglGo3HIz2/YsAGpqalDPrzBzHQt5yUTEY1Rpd4IjdofIUEKsUshIh/ltpCsUqnQ3t4+eCHJ4KVCQkIQExODpKQkBAQEIDMzEzU1NUN+vrCwEGVlZUM+vMHMdC10ehNa281il0JE5LUq9SaksB+ZiETktpCck5OD3bt3w263o7i4GPHx8QOPKRQKaLVaNDc3w263o6KiAjExMe4qxaNiI4MQGRaAfVxNJiIaNU62ICKxuW3jnlqtxsKFC7F48WLIZDKsWbMG27ZtQ2xsLPLz8/Hggw/ivvvuQ29vL66++mpoNBp3leJRgiBgZlok9pQaMD8//sd/gIiIhnA6ndDpjbjuohSxSyEiH+a2kAwABQUFKCgoGLh/8mpyVlYW3njjDXdeXjQz07VY+8Ze2OwOyKQcRU1ENBLHW7rRZbZx/BsRiYoJzg2yUjTotTlQerRV7FKIiLxOpd6I8BAlQoOVYpdCRD6MIdkNlHIZslI02MtRcEREI6bTGzn6jYhEx5DsJjPTtJyXTEQ0CpV6I5InsdWCiMTFkOwmeemROHa8A4a2brFLISLyGn2b9kxIjuNKMhGJiyHZTWI0QZgUEYjiqhaxSyEi8hqNrd3o7OlluwURiY4h2Y1umj8VkaEBeP/rKhxtaIfZYhO7JCKicU2nNyFMpUCYipv2iEhcbh0B5+sma1VYsf4LOJyARADWr5iHhGiV2GUREY1blXojkiZxFZmIxMeVZDcqPdYKh7Pvc4cTOKxrFrcgIqJxrpKTLYhonGBIdqOsZA0kQt/nEgHISAwXtyAionHsxKa9FB4iQl7KbLOgxliHf5d/hhpjHcw2i09ce6Jiu4UbRYUFYP2KeTika0ZKrBof7zyKC6bHIjOJYZmI6FRNbT3o6LYihZMtaJxzOp3otffCYrfCYrPCYrfC5rDD7rBj1X+egtPphCAIeGL+Q2i3dMBis8IJJ5xO5+CtE6d/bcgt4ITjLN83+LhSpkBS6GQ88unTA9d+esFvMFk9Sex/VV6NIdmNlAoZEqJVSIhWwel0YnfJcfz2xW+xYnEezsuOEbs8IqJxpVJvhDqYm/YmArPNAkNnM4oN5ciMnIrIIA2UMoXbrtcXOvvCo8VmRVN3K0oM5UiPmIJAuT96es0w2yyw2q0w26yn3PZ93WIbDLyDtxZY7FZYbVaYT7l1wjmkhoyIKcjSpsHpdA7UdPB4CY4a9dC1HIUgCBAg/MAtIAiS/tthPH7K9ySExqLH1jPk2nvqi1DSVIGYYC3SI1LgJ/Vz27//iYoh2UMEQcDSKzIQofbH06/vQes103D1BUlil0VENG6cOEREEASxS6FRcDgc0Lc3oNZUD01gGFZ/um5gVXP1Rfdjc9G7aOsxwgmcZfXU2f+4Aw44ASf6b50/cIshYTXcPxS/mL0Uj3/x3MC1H5l7L57f+RpazUYopQrIZXIopfIhtwqZAgqpHAqZHEqZAiHK4IH7Cqmi71Ym7/+a4qTHBm/lMjn0pgZsKf7XwLVnTsrB9RmXe+Tff42xDpsPvTdw7eyodHxc8SXeOLgdADBNm4rc6ExMj56GyEC+oz0cDMkedvmcRISplPjTpr1oNvbgtiszIJHwBYGISKc3YQpbLbyC0+lEU3crKluOorL1KHStx1DVVgOLzYIL42cjLiR6yKpmdVsNCrKvQYel60dXUAEBkv4/lPpuhR+4BSSCZMj3Bfgpsbf+8JBr600N2HDlHyCVSN3+B1hkkAZPL/gNSgwVyNBO9WgYPdO1fzF7KZbPLEBpUyUONBTj3+Wf469738Kk4Kj+wJzJVeazYEgWwexp0Vhz9xz84a870Wzqwf2LcuEnk4pdFhGRaJxOJ3R1Rlx2boLYpdAZtFs6oesPwyeCcbulE6H+IUgJS8D0qAzcmHkFkkInI1AegBpjHd489O5Jq5oZmBzimTbDLG0qBEEYuHamNhUyqWfijlKmwGT1JFF6gX/o2n5SP2RHpSM7Kh1Lc2+EoasFBxoOY19DMT75+itAEDAtcipyo6chNzoTEVxlHiA4T/y55QVSU1NRVlYmdhkuU9fUiUdf+g7asAA8vCwfgf78S47Il3jrc5o76m5q68Htj3+MVx5ZgIhQf5f+bhoZi82K6rZaVLb2rxK3HEVjVzP8/ZRICYtHclgCUvo/wgLOvPJ/oif55FVNd/Ykj5drexurvRdHmiqxv6EY+xsOo76jEZNUUQOBOU2TPOFXmc/2fMaQLLK2DjP+8NfvYbM78eid50Cj5osDka/w1HPali1bsG3bNvj5+eGJJ55AXFzcwGPbt2/Hpk2bIJFIcNVVV+G222770d/njrq/O9SA//3HAbz+u8vYk+xBdocd+vaG/tXhY6hsPYpaUz0kggQJ6ti+MByegOSweEQHRw60N9DEZOhs7gvMx4tR3FgGCAKyIlMxPTpzwq4yMySPcz0WG/742m4ca2jH75afi3ieykfkEzzxnGY0GrF8+XJs3rwZJSUlePnll7F+/fqBxy+77DJs27YNSqUSV111Fd555x3I5XKP173p36WoqDXi93ed69Lf68tOnTAREahBh6Wjb4W4v2Wiuq0WVnsvYlTagdXh5LB4xKsnTfgVRDo7q70XpU0V2N9QjAMNxajvaESsKhrTozMxIzoTaZoUj7WxuNPZns+8/59uAvBXyPDI7bPx/D8O4qGNX+E3t89GVrJG7LKIaAIoKipCfn4+ZDIZsrOzUV1dPeTxpKQkdHd3AwCUSiWkUnH2R1TqjUjmISIu1djZjAc/XjPQm/vbuffhf3f+HQCQHB6PGTFZuGnaVUgKjUeAnO9iejuH1YxeYyPMx4qhjM+En1oLiXz04xTlUj/kRGUgJyoDyP0pGjubBgLzx5VfQCJIME2bhtyoTMyInoau3m6PjfzzFIbkcUImlaDwpunQqP2x+sXv8OuCGbggl0PAiWhsTCYTQkIGw+epbx5eeumlWLhwIaRSKe64447TQvKGDRuwceNGt9Z44qS9BbPj3XodX9Bt7cG+/t5SqSAZMuXhqFGPP166CipFsMhVkqs5ei3obWtA3csPAk4HIEgw6c61UES67v8pbVAELptyES6bchGs9l6UGCpwoOEwvq7ZhajgiCFj956Y/xD8ZQqEBYR6dVhmSB5HBEHALZemITzEH89s3ouW9h4snJsidllENA48+uijWLhwIXJzc0f0cyqVashbiRLJYE9pZ2cnXnzxRezYsQNyuRy333475s+fj5iYwSkEhYWFKCwsHPI7U1NTR/lPcWat7WYYOy1IieX4t9Ew9piwu64Iu+sO4JChDEHyQFyUcA5mxs0YMrM3OyqdAdmLOB122Ls7YO8ywt7Z1vfRZYSt0zjwub3TCFtnGxRRSfBPzOoLyH0/jO7KfbB3tEKuTYQsyLX/b8mlfpgenYHp0RkAgA/LPh3yB9mBhmIUNR5BaVMFQhTBiAwMR0RgOCKDNIgMDEdkYN+tJiBsXLdsjN/KfNil58QjTKXAH1/fgyZjD+64ehpnKRP5uAsvvBCvvvoqVq9ejUsvvRTXXnvtkA14PyQnJwfPP/887HY7jhw5gvj4wZUliUQCPz8/BAQEQCKRQKlUorOz053/GGdUWWtEcICcUy1G4HiHAbvqDmK3/gDKW6oRGaRB/qQcXJ9xBaaGJ0IikcBss4g2s1dMrm47cPW1HVbzQPAdCLxnCsBdpoHQK/EPgjRQDVlQKKRBoZCpNFDETIE0SA1ZoBpSlQZOmxVtX24ZWEn2n5yB5k9ehbW+AvKoZAQk5yIgJReKmCkQJK5tq5p2yti9WbHTceXUi2Hoaun/aIahqwX69uPYX38Yhq4W9NjMEAQBYf7qwRAdeFKIDgpHmFI95A97T+PGvXGsvKYNf3j5e0xL1uDXBTMg9+MsZaKJZDTPae3t7fjggw/wf//3f4iNjcX111+Pa6655qyb7TZv3ox3330XMpkMa9aswd69exEbG4v8/Hz87W9/wwcffABBEDB9+nQ8/PDDbqn7bN7YcQRHjrXisZ/PcdnvnGhOtEvs0h/ArroDqDXVI1Edh1mxOcifNB1xITGcCoK+1VeroQZ1rwy2HcQsexL2LhOcNqtbry3I5JAGhqD+b6sGr730cXQc+hy9TbX9IbgNTqu57weksr7QG6iGNEjdF34DQwc+7wvFakgD1RBkP76J8uSA7h8/DTJ1JAQ/BXqba9Gt249u3X6Ya0ohkSvhn5SDgORc+CflumSVeaRj95xOJ7qs3QPh2dDVDEPnYJhu6mpBr8MGqUSKiICwoQE6aPBzlSIYFrt1TEegc7qFF2to7sKjf/kOYSolHvlZPoICzr7rnIi8x0if0+rr6/Huu+/io48+QnJyMq644gocOHAAe/bswebNm91Y6VCufi7+/V+/R0K0CrddmeGy3zkR2B12lDXrsEt/ALvrDqK5pw3pmhTMmpSDWbHTfWZl+FROhx02owG9rQ3obTvef9uA3tYGyEIi4B+fibYv3hr4/tCLCmBtrIGlvsKtdSkmTYE8YjLavhj8fzF03q2QKPzhtFlPC8ASZaDH/7BxWHrQc/QQunX70aPbB1t7M+RRSX2rzMkzoJjk+lXmUdXpdMBobkdTV8uQ8HzitqW7DQ6nA9FBkbhr1mL84fM/D6xiP73gNyM6zIXTLbxYtCYQTxdegD+8/D0e3PgVfrf8XESGBohdFhF52K233gq73Y7rrrsOmzZtQlBQEADgkksuwfLly0Wubmyq6oyYP2uy2GWMC1Z7L4qOl2JX3QHsrT8Ec68ZWVHpuCHzSsyMyYJK6Rs9xU67DTZT0xmDsM3UBDjskCgD4RcaDb+waCiiUxCUeQH8IuIgCJIhbQeBU/IRet6NHqnbYjiGti/fHrx2Sh7kkePnv22Jwh+BqfkITM2H0+lEb7N+IDAbv3u3b5U5MbtvlTk5F7KgUHHqFCQI81cjzF+NVE3yaY/bHXa0dLehy9qNw03lQ/qhSwwVLjvxkCvJXsJsseFPm/ZApzfhd8vPQWIMRyURebuRPKdVVVUhKSnJzRUNjyufi1vbzbjt9x/hLw/PR1R4oEt+p7fpsnZjX/1h7K47iP3HiyEVJJgRPQ35sdMxPSoDSj/P9NN62tAg3IDe1sEwPBiEg+AXFg2/0CjI+m/77kdDGnDmPxjO1HYgRk+yp689Vg5rD3qOHkZ35b7BVWZtYn8vc964WWU+VY2xDitPGnX49KWPjOgIdBdsPWkAACAASURBVK4kTwBKhQy/WZaPF7YfwkMbv8ZvluUjZ2qE2GURkYe8+uqrWLlyJVSqvsOGTCYT1q1bhz/84Q8iVzY2lXojgvz9oA3zrXfIWnuM2FN3ELvrDuJwYxlUimDMnJSNB867C5kRU8f1jv/hGLKBLS4dTocNlrryoUHYaACcjsEgHBYNxaQpCMq6cCAMS/1HvnIukSuhiIx36fgzb7j2WEnk/gicOguBU2edvsq88z1I/BTwTzyplzlYnFXmU0UGady2QdW7/y/0MVKpBL+4IRsatRK/++t3uO/mXFyU9+O724nI+x06dGggIANASEgIioqKRKzINXS1fYeITMRNZ6eeeKeQKfB97T7sqjuAipZqRAdFYlbsdPw08yqkhCdMiCOfnbZemOsqIEilqH/tkYG2g+hbVqOnuggSRYBLgjC5lyAIkEfEQR4RB/U51wyuMuv2oe2rLWj61/+etMo8A36R8bCZmkSZKKKUKTBZPcllLRYnY0j2MoIg4Ob5qQhX+WP92/vRbDLjhnkpE/IFhoiGamlpQXh43ypJU1PTaQeDeKNKvWnCzkc2dDYPeRv4kbn3osRQjryYLNw9awkmqaImxHN3r7ER3ZV9K449xw5DGZcOZXzmkJm91qZaRN20StxCadROW2VuqUO3bh96dPvRUfINIq+8Gw1v/mFwqsfPngJsvYBEColcCYnCH4KfEhK5EoIXvUviPZXSEPPzJyNMpcRTr+1Cs7EHyxdmQcpZykQTVmFhIQoKCjBnzhw4nU58//33WLXK+0OHrs6Ii/JixS7D5ax2Kw4cLxmyoajW1IBVc+8RubKxc9isMNeUDLwV39tSDz9NLAKSc6HKvxL+cRmwttaj7fPNgzN746eJXTa5iCAIkGtiIdfEQj37Gjh6LWjf9/GQP4p6dPthrj2CnqoDAE75Y14q6wvOfkoICn9IToRnuRISeX+YVigh8fPv/1r/Y35Dw/aJr0Migc3U7JZVbIZkLzYjLRJP/OJ8/P6v36O13YwVi/Og4CxlognpkksuwfTp03Hw4EEAwH333YewsDCRqxqbtg4zWkzmCbWSbHPY8VnVt/iP7isszb1xyAEL07SuPanQk3rbjveH4v3oOXYYgAD/hCyEzLoK/sm58FNHDvl+P7UWk+5cO2QDG01Mfb3K2YAgGZzqMTUfoeffCKfTCafNCqfVDIe1Bw6refDz3hOfm+Hsf+zE47b2FjgsPXD2mk/6GfPAz5wI5NLgcERefQ8aNj/mluO4hxWSd+/ejczMTAQEBGD79u04cuQIbr311mGd9kTulRKrxtOFF+B3f/kev33hWzxy+2yoAjlLmWgikkqlCA8Ph9VqhU6ng06nw6xZs8Qua9R0ehMClTJEhXv/pj2n04nddQfxRtF2dFm7cWPmlUhUx3ntiXcOmxXmY8UDwbi39cRq8QyE5F8FZVz6WQ+48OYNbDRyP/RHkSAIEPwUgJ8C0kDXTOU6OXg7HXZ0ln47ZBXbfKzYsyH5sccew3vvvYeioiJs3rwZt9xyCx566CG8+eabLimCxiYqPBB/KrwAj738PR7c8BV+t/wcnx2lRDRRvfnmm9i2bRtqa2uRm5uLXbt2IS8vz6tDcqXeiORYtdf35R5p0uGNg9tw1KjHVanzcXXafAT49R2xPVnung1F7tDbdhzdlfv6TmY7dhiQSPpWi2dfDf/k6fAL4WownZkn/yg6OXgDgH9C1pBVbFe29gwrJJ94Avvkk09wyy23YOHChfjb3/7msiJo7FSBcjx+93lYu2kPVm74Co/eec6EeguTyNdt3rwZ27dvx/XXX48XXngBer0ejz32mNhljUllbV9I9lb69ga8WfQu9tUfwsVJ5+HX592FUH/vmWHv6LX09xbv618tboBfRBwCknOhPucaKOPSIEh//DhkIjG5s7VnWCE5IiICq1atws6dO7Ft2zZYrVbY7XaXFUGuofCT4v/dlo+/vHMIH39/FJidgNJjrchK1iAqLABKBVvQibyVQqGATCaDVCpFT08PYmNjUVNTI3ZZY6KrM+HCXO9YZT1Za48RWw9/gM+qv8XMmGw8c9lvEaOKErusYeltbUC3bj+6dftgPlbcv1qcjZDZ1yAgOReyEM7fJ+/izlXsYaWm5557Dl9//TV+/vOfQ61Wo6mpCQ899JDLi6Gxk0oE/Py6LJTXtGHF+i/gcAISAVi/Yh4SolU//guIaFzKyMhAe3s7brjhBtx0000ICgpCVlaW2GWNmqnTgmZjj1e949Xd24P3jnyMf5X9F4mhk/H7i399xiNzx4u+Az2Oo6f6EJSTpqKrfBdM370Dv4jJ/avF13K1mOgshhWSi4uLcf7555+2cY/GJ0EQUF5rhKN/6orDCRzWNTMkE3kpp9OJu+++GyqVCrfeeivmzp2Lrq4upKWliV3aqFXqjfBXyLxi/4TNbsPHui/xz5J/QyUPwn3n3oGZMdnjspfaYe6Cua4c1sajUMSmomHTo4Oza5c9gZBZV0IW7N1TUYg8hRv3JqisZA0kAgZWkrOSNWKXRESjJAgC7rrrLrz//vsAMCEmC/Vt2guBZBzPd3c4Hfiudi/eKnoPVnsvCrKuxbzEcyGVjJ9Rm7bONphrS2GuKYW5thRWwzEIciVUuQtgqasYsuvfUlcB5awp4hZM5EW4cW+CigoLwPoV81BU0YSEGBU6e6xil0REY5CXl4dvv/0Wc+bMEbsUl9CN85P2DjcewaaD29HQYcC16QtwxdSLoZQpRK3J6XTC1taAnv5AbK4tha3tOKSBaijj0hGcMw/KuHTII+MhSKSwGI4Bn23igR5Eo8SNexOUUiFDQrQKCdEqfLTzKP72fgk2rpyH8BB/sUsjolH45JNP8NZbbyEoKAhK5eBpUl9//bWIVY2eTm/EnOwYscs4zTGjHm8c3I5DhjIsSL4QD194D1TKYFFqcTrssBqODVkptncZIQuNgjIuA6Hn3QBlXDpkoWc+3poHehCNzag27hkMBm7c8yI/mRWPL/fV4bm3D+B3y88Zl310RHR233zzjdgluEx7lxWGth6kxI6fcWnNXa146/B7+OrYLpwbl4dnL38UUUGenfTgsFlhqa/oD8RHYK4rg9PSA7k2Acq4dIRfegeUsemQBYcO6/fxQA+isRlWSA4ICIBGo8F//vMfAMCMGTNw/vnnu7Uwch2JRMB9i3JRuPYz7PjuKC6fkyh2SUQ0Qj+0YuyNz8V9m/akiNEEiV0KOi1d2F66AzsqPkeqJhlPzH8IyWGeCZV2cxcs+iMw15aip6YUloZKCBCgiEmBMi4dIflXQjlpKiTK8b+5kWgiGlZI3rhxI7755htcccUVAIB169Zhzpw5+OUvf+nW4sh1IkMD8PPrsvD8P4uQMyUCMRHivzgR0fB98MEHA59brVbs2bMH6enpXhmSdXojkiapRd20Z7X3YkfFZ9hesgOagDCsPP9/kBOV4dJ32vpGsDXCfKwYyvhMSP2D+1snSvo32dVAUPhDGZuKgJQ8hF98K+TRyZDI5C6rgYhGb1gh+aOPPsL27dshk/V9e0FBAa677jqGZC8zLy8O3x8+jmc278Mff3k+pFKJ2CUR0TA9+eSTQ+63t7djxYoVIlUzNjq9CckitVo4HA58eWwn3j7cNynkttyf4sL42ZBIXP982Nt2HHUvrxzYOBdd8FsYv3sHfqFRCJ4+v3+T3WQI42haBhENGvZ0i5aWFmi1WgBAW1ubW55QyL0EQcAvb8zBPWs/wz8+q8DN81PFLomIRkkul+PYsWNilzEqlXojZk/zzIxns80CQ2czig3lSNUkYVfdQXxU8Tmuy7gMl02ZB7kLD9JwOh2wNFShp3IfbF3Gvt7hk0ewNdVi0u1/4r4QIi8xrJC8YsUKFBQUYMqUKXA6naiqqsLq1avdXRu5QUiQAvfeNB1P/G0X8tK043oEExENuvnmmwfCld1uR2NjI5YuXSpyVSPX0W1FY2u3x557DJ3NWPnxGjidTgiCgMcufgBXT52PQEWAS36/vbsdPVUH0a3bh+6qA3D0dPa1T6TPgX9cGtq+3DKwkhyQkMWATORFhhWS586di3//+9+orq4GACQmJmLnzp1uLYzcZ1ZGFC6ZNRnPvLkPf/7VXMj9+FYf0Xj3zDPPDHwulUoRHh4OPz/vO05YpzdCKZd6bF9EsaEcTmff8aNOpxNVrTWYqkka9e9zOh2w1OvQo9uPbt0+WOorIQ1Swz8pF5rLlsM/MQfS/o12DquZI9iIvNiwQjIAKBSKIUegrl69Gp9//rk7aiIPuOOaabh33Wd47cNS3HktB8wTjXcVFRXIy8tDcHDfzN6Ojg58++23mDt3rsiVjUyl3oTEmBBIPbRpLzNyKgRBGFhJztBOHfHvsHe3o7vqQF8wPnm1eGo+NJf/HHJtwhlXiDmCjci7DTskn+rEX+bknfwVMvyqYAZWPf8N8jO1yE7x7DxQIhqZZ599Fu++++7A/eDgYDz77LNeF5J1eiNS4jzX5hURqMHqi+5DdVstsqMyEBkY/qM/43TYYWnQoVu3Hz26/QOrxQHJudBcdhf8E7MHVouJaOIadUhmX5X3y0gMx/UXpeDPb+3HhhXzEOjvfW/dEvkKh8Nx2tdsNpsIlYyNTm/CogVaj13PZDZh4/d/x58ufRjBih9u8RhYLa7ch+7qg32rxXFpCEzNh+aK/+k76pmve0Q+5awh+WzzN9va2lxeDHneLZemYk9pI1565xB+VTBD7HKI6Afk5+fj4YcfxqJFiwAAmzdvxuzZs0WuamQ6e3rR0NKFZA9uGC42lEMhk58WkM+8Whzat1p8+V0ISMjmIR5EPu6sIfmHTniiicNPJsWKxXn41bNf4JxpUTg3K0bskojoDFatWoW3334bf/nLXwAA5557Lm6++WaRqxoZnd4IuZ8UsR48zKi48QjSgqNh2v0hlJMzYO/pROeB//T1Flu6oYxNQ2DqbK4WE9FpRt1uQRNHQrQKSy5Pw8atB5EWH4ZQlVLskojoFC0tLbjxxhuxePFiAIDFYkFzc/PA/HpvoNMbkRSj8thBRk6nE8WGclxaU4eWDnPfgR63PAppUBgirvgf+CdkcbWYiH4QTwQhAMC1c1MQpw3Ghq0HuCmTaBz6xS9+MWSVUxAErzv1VKc3eXQ2e2NXM9osHUjqtvZ9wemAtakG4fOXIjDtHAZkIjort4bkLVu2YNGiRViyZAlqa2tPe7yjowOzZ8/Gjh073FkGDYNUIuD+Rbk4rGvGxztrxC6HiE5hs9kgl8sH7svlclitVhErGjmJREBG0o9Pl3CVEkM5ovzDoDqx51GQwD+eIy+JaHjcFpKNRiO2bt2KTZs2YeXKlVi7du1p3/Pyyy8jJyfHXSXQCEWFB2L5tVl4+b1DON7SJXY5RHSSqKgovPPOOwP3t2/fjqioKBErGj6zxYbqehMmRQYhJiIQZotnpnIUG8qRqZ2K6FtWI2z+MsTeuY4HehDRsLktJBcVFSE/Px8ymQzZ2dkDp/Wd0NzcjNraWmRlZbmrBBqF+fmTkZ0SgWfe3Ae7g20XROPFY489hk8//RRz587F3Llz8eWXX+Kqq64Su6xhOd7ajfuf+Rxv7DiCXz/7BY63drv9mk6nEyWGCqT6R8Dw3kYET78E8sjJkMi554KIhsdtIdlkMiEkJGTg/ql9ri+88AKWL1/+gz+/YcMGpKamDvkg9xMEAff8dDrqmzvxzueVYpdDRP0iIyPx9NNP4+GHH0Zubi727dvnNROIDumaceJvbocTOKxrdvs1G7ua0dLThsTObkgDgiFVBLj9mkQ0sbhtuoVKpUJZWdnAfYlkMI/X1taivb0daWlp+OSTT87484WFhSgsLBzyNQZlz1AHK3DPT6fjj6/twYy0SCTGhPz4DxGRW1itVnz++ef497//jaKiIpx77rnYs2cPvvjiC0ilUrHLG5asZA0kQl9Algh9992txFCB6KBIKOp08IvPdPv1iGjicVtIzsnJwfPPPw+73Y4jR44gPn7w7PrS0lLU1NTgjjvuQE1NDQIDA5GcnIwpU6a4qxwaoXOmRWPujEl45s19eOb+C+En844XY6KJZvbs2UhLS8O9996LtWvXQiqV4uKLL/aagAwAUWEBWL9iHg7rmpGVrIE2zP2ruiWGcmREToH5y08QnHOx269HRBOP20KyWq3GwoULsXjxYshkMqxZswbbtm1DbGwsFixYgAULFgDoa6uYMmUKA/I4dNfCLBSu/Qxv7DiCZVdxJYZIDA888AB27NiBxx9/HAsWLMBll13mdQdeKBUyJESrkBCt8sj1nE4nipvKcVP8+XCYu6CcnOGR6xLRxCI4vWgobmpq6pAWDnK/Q7pmPPLCt3ji7vOQ6cHRTUS+YCTPac3Nzfjoo4+wY8cOHDx4EEuWLMH8+fORm5vr5ipPN96fixs7m1D4wWr8Ke4SyEp3IvaOp8UuiYjGqbM9n/EwETqrrGQNrrkgCc9u3oduc6/Y5RD5LI1Gg8WLF+P111/Hp59+ipiYGDzzzDNilzUuFff3IyvrdPBnPzIRjRJDMv2oJZenQymX4q/vHha7FCLC0MBMpysxlCM9IgXmmhIoJzMkE9HoMCTTj5L7SfHrW/Lw2d5a7DzcIHY5REQ/yOl0oqSpAlOVYexHJqIxYUimYUmaFIJbLk3Dxq0HYeq0iF0OEdEZNXW1oLm7FYmd3ZBHJUKqDBS7JCLyUgzJNGzXz5uCaE0g/vcfB087HIaIaDwoNpQjKigC/uxHJqIxYkimYZNKBPyqYAb2lxnw6Z5ascshIjpNcVM5MiKmwFxTyn5kIhoThmQakWhNIG6/Zhpe3H4IhtZuscshIhqixFCBqYpQ9iMT0ZgxJNOIXXZOPDKTwvHnt/bD4WDbBRGND4YT/chdPexHJqIxY0imERMEAYU3TcfRhna895VO7HKIiAD0jX7TDvQjTxO7HCLycgzJNCphKiV+eWMOXvuwFMeOt4tdDhERig0n9SNz0x4RjRFDMo3aeTkxOC8nBs+8sQ+9NofY5RCRjysxlCNVEQqHpRvKuHSxyyEiL8eQTGPy8+uy0d5lwVufnPnccyIiTzB0taDpRD+ylv3IRDR2DMk0JkH+frh/0Qz889MKHDnaKnY5ROSjSgzl0AZqOB+ZiFyGIZnGLGdqBK48LxHPbN4Hs8UmdjlEdIotW7Zg0aJFWLJkCWprh844v/3227FkyRIsWbIE06ZNg8lkEqnKsSkxVPT3I5ewH5mIXIIhmVxi6ZUZkEkFvPJ+sdilENFJjEYjtm7dik2bNmHlypVYu3btkMdfeeUVvP7661i9ejXy8/MREhIiUqVjU9xUjqnKUDgsPexHJiKXYEgml1D4SfHrgjx8vPMY9pQ2il0OEfUrKipCfn4+ZDIZsrOzUV1dfcbv+9e//oUrr7zSw9W5hqGrBU1dLUjoZD8yEbkOQzK5TEqcGjf/JBXPvb0fHd1WscshIgAmk2nI6rDTeeYDgD755BMsWLDgtK9v2LABqampQz7GmxP9yIF1VexHJiKXkYldAE0sN10yBYADtY0d0NWZkJWsQVRYAJQK/qdGJAaVSoWyssHpMxLJ6WsjBw4cQGJiIoKDg097rLCwEIWFhUO+Nt6CcomhAukRKTB//V8E584XuxwimiC4kkwuJZVKMDszBqv+92u8tP0Q7lv3GY63dotdFpHPysnJwe7du2G321FcXIz4+PjTvudf//oXrrrqKhGqc42SphPzkXvgz35kInIRLu+RyxVXt8DR/46uwwkc1jUjIVolblFEPkqtVmPhwoVYvHgxZDIZ1qxZg23btiE2Nhb5+fmw2+348ssv8cADD4hd6qg0dbXA0NWChC4zFFGJkLAfmYhchCGZXC4rWQOJ0BeQJQKQmRQudklEPq2goAAFBQUD909eTZZKpfj444/FKMslSgwViAwMR2BdFfzYj0xELsSQTC4XFRaA9Svm4ZCuGcmTQvDdoXpEawKhlPM/NyJyreKm8r75yF//F6rcn4hdDhFNIOxJJpdTKmRIiFbh6vOTEB+twpf767Hh7QM/uKueiGi0SgzlmKo4MR85TexyiGgCYUgmtwpU+uG3d8zG3jIDtv63QuxyiGgCae5q7etH7uxmPzIRuRxDMrndpIggPLhkJt786Ai+P9wgdjlENEGUNFUgIjAcQfXVPIqaiFyOIZk8YkZqJG6/OhPPvLkXRxvaxS6HiCaAYkM5MiJSYK4pgf/kaWKXQ0QTDEMyeczVFyTh/JxJePyVnTB1WsQuh4i8XImhHKnyMDisZvYjE5HLMSSTxwiCgLtvyEaYSok/vrYHNrtD7JKIyEs1d7WisasZCV3sRyYi92BIJo/yk0mxatksNLR04aV3DoldDhF5qZKmCkQEhLEfmYjchiGZPC40WInf3j4bn+6pxYffVotdDhF5oRJD/3zkmhL4x7MfmYhcjyGZRJE0KQS/WjQDf3nnEIoqm8Quh4i8THFTBaYqTvQjp4tdDhFNQAzJJJrzcmLw00um4qm/78bxli6xyyEiL9Hc3YrGziYkdnVDEZUEiSJA7JKIaAJiSCZRLfpJKqYla/DYKzvRbe4Vuxwi8gIlhgpo2I9MRG7GkEyikkgE/KpgBiSCgGfe3AeHg0dXE9HZlZw8H5khmYjchCGZROevkOG3t8/GkWOt2LSjVOxyiGicK2mqQKoilP3IRORWDMk0LkSGBWDVbfnY/nklvtinF7scIhqnWrrbcLyzCYmdZvYjE5FbMSTTuJGZFI7/uT4Hz729HxW1bWKXQ0Tj0Il+5MB6HfuRicitGJJpXLn0nHgsOCcea17dhdZ2s9jlENE4U9zU349ce4T9yETkVgzJNO7cec00xEUG44lXd8Haaxe7HCIaR0oM5ZiqCIWT/chE5GYMyTTuSKUSPLh0Jtq7rdi49QCcTk68ICKgtds42I8cncx+ZCJyK4ZkGpeCA+T47e2z8f3h49j+uU7scohoHChpKkd4QCiC2I9MRB7AkEzjVpw2GA8umYnXPizBntJGscshIpEVGyqQETGlrx95MkMyEbkXQzKNazPTtVh6RTqe3rQHtY0dYpdDRCIqMZQjdaAfOU3scohogmNIpnHvuotSkJ8Zhcde2YnObqvY5RCRCFq7jWjoNCChs4f9yETkEQzJNO4JgoDCn05HcIAf/vjaHtjtDrFLIiIPK2kqR7h/KILrq9iPTEQewZBMXkHuJ8XDy/JR09iBV94vFrscIvKwEkPF4Hxk9iMTkQcwJJPXCA/xx29+lo8d3x3FxzuPiV0OEXlQcdPJ85HZj0xE7seQTF5l6uRQFN6ci//7ZxGKq1rELoeIPKC1x4iGDgMSuzgfmYg8hyGZvM5FM2KxcG4ynvz7LhjausUuh4jcrMRQgTB/NfuRicijGJLJKy25PB2pk8Pw+Cs7YbbYxC6HiNyoxFDO+chE5HEMyeSVJBIBKxbPgN3hxLNv7YPDwaOriSaqkqYKpCrU/f3I6WKXQ0Q+giGZvFaA0g+P/Gw2DlU24+1PysQuh4jcoK3HhPqOxpP6kf3FLomIfARDMnm1aE0gHlo6C1v+W45viurFLoeIXKykqby/H7ma/chE5FEMyeT1cqZE4M5rs/Ds5n2oqjOJXQ4RuVCxoQLpJ+Yjx08Tuxwi8iFuDclbtmzBokWLsGTJEtTW1g58vb29HbfddhtuueUWFBQUoLiYh0PQ2FwxJwHz8uLw+Ks7Yeq0iF0OEblIiaEcU+X985FjOR+ZiDxH5q5fbDQasXXrVmzevBklJSVYu3Yt1q9fDwCQy+X405/+BK1WC51Oh8cffxyvvvqqu0ohHyAIAu5amIXtnytR29iBLw/UIStZg6iwACgVbvvPnIjc6EQ/clJQD/uRicjj3JYeioqKkJ+fD5lMhuzsbFRXVw88plQqoVQqAfQFZqlUetrPb9iwARs3bnRXeTQB+ckkmJURhfuf+RwOJyARgPUr5iEhWiV2aUQ0CqVNFQj1D0FwfTUU7EcmIg9zW7uFyWRCSEjIwH2n8/QRXU6nE08++STuvPPO0x4rLCxEWVnZkA+iH3O4qgUnpsE5nMChyiZxCyKiUSvun49sYT8yEYnAbSFZpVKhvb198EKS0y+1Zs0a5Ofn45xzznFXGeRjspI1kAh9n0sEIDEmhKfyEXmpEkNFXz9yr4X9yETkcW4LyTk5Odi9ezfsdjuKi4sRHx8/5PEXXngBUqkUy5Ytc1cJ5IOiwgKwfsU8/Py6LKxfMQ9VdSb8+s9foLymTezSiGgEjD0m1HUcR1IX+5GJSBxu60lWq9VYuHAhFi9eDJlMhjVr1mDbtm2IjY1FXFwc1q9fj7y8PCxZsgSRkZFYt26du0ohH6JUyJAQrRroQ56sDYahrQernv8GD96ah9nTokWukIiGo6SpAqHKvn5kZQJbLYjI89y67b+goAAFBQUD909eTS4tLXXnpYkA9B1ffee10xAZ5o8n/74by6+dhivPTxK7LCL6ESWGCmREpMDyzedQz7pS7HKIyAdxNhb5hGsuSEaE2h9rN+1FY1sPll2ZAcmJ5mUiGneKm8qxIDKrrx85jv3IROR5PHGPfMa5WTFY84vz8N/dNXh60x5Ye+1il0REZ2A0t6Ouvb8fOSYFEjn7kYnI8xiSyaekxYfh6XsvgK7OhN+++C06uq1il0REpygxVECtVCG4vgr+nI9MRCJhSCafE6MJwtOFF8DhcGLlc1/heEuX2CUR0UlKTpqPrJzMkExE4mBIJp8UEqTA43efh8lRwVj53FccEUc0jpQ0nZiPbGU/MhGJhiGZfJbCT4qHls7ChTMm4eH/+wa7io+LXRKRzzOZ26Fvb0BSt5n9yEQkKoZk8mlSiYDl12ZhyeXpeOJvu/DBN9Vil0Tk00qa+vqRVexHJiKRcQQcEYBrL0yGRu2PZ97Yi6a2biy9giPiiMRQNv2FDgAAHs1JREFUbChH+sB85KvELoeIfBhDMlG/87JjEBasxGOv7IShrQf3L8qF3E8qdllEPqXEUIH5ERlw2nrZj0xEomK7BdFJ0hPDsPbeC1BZa8Tql77jiDgiDzrRj5zYbYYimv3IRCQuhmSiU8REBOHpey+AzebAgxs4Io6835YtW7Bo0SIsWbIEtbW1Qx5raWnBvffei6VLl2LFihUiVdintKkSIUoVQuqr2Y9MRKJjSCY6g74RcXMQGxmElRu+QkUtR8SRdzIajdi6dSs2bdqElStXYu3atUMef+qpp/DAAw/gtddew7p160Sqsk+xoRwZESl985EZkolIZAzJRD9AKZfh/92WjwumT8Kq57/BrhKOiCPvU1RUhPz8fMhkMmRnZ6O6enCCi91uR1VVFTZs2IBbb70VH374oYiV9h0iMlWu7utHjk0VtRYiIm7cIzoLqUTAXQuzEBkagCde3YWfX5eFy+ckil0W0bCZTCaEhIQM3Hc6nQOft7S0oKysDE8//TSioqJwyy23YM6cOVCr1QPfs2HDBmzcuNH9dZrbUdvegKSARPYjE9G4wJBMNAwL5yYjQu2PdW/uRWMrR8SR91CpVCgrKxu4L5EMvoEYEhKCmJgYJCUlAQAyMzNRU1MzJCQXFhaisLBwyO9MTXX9Km9pUyVCFMEIqT/KVgsiGhfYbkE0TOflxODx/5mDj3fWYN2be9Frs4tdEtGPysnJwe7du2G321FcXIz4+PiBxxQKBbRaLZqbm2G321FRUYGYmBhR6iwxVPT1I+vZj0xE4wNDMtEIZCSG4+l7L0B5TRtWv/QdOjkijsY5tVqNhQsXYvHixXjyySexYsUKbNu2Dbt27QIAPPjgg7jvvvtQUFCAq6++GhqNRpQ6i5vKMVURyn5kIho3BOfJDWrjXGpq6pC3DYnEYuyw4LFXvkePxYZH7zwX2rAAsUsiL+Stz2murrvd3IE7330Qv4+aA1V1KSYte8Jlv5uI6GzO9nzGlWSiUVAHK7Dm7vMQownCyue+RGWtUeySiLxWSVPFQD8y5yMT0XjBkEw0Skq5DKuW5eO87Bisev5r7CltFLskIq9UYqhAOvuRiWic4XQLojGQSgTcdV0WIsP+f3v3Hhd1mfcN/DMnmGEGGIaTghwE5KAI6nqI7ASkafoytVtTWbUe17UsbQ1r8e5eU8tjPG3smq8tsb13LTMre7rTtPXwqOtqUmuPKAIqIsiQgMpwZoCZ6/kDHeUoCDgz8Hm/Xr7kd/4O13DNl4vv73c54UR6IVzUDsjOL8XQYA/00zlB6cgfMaJ7OV9yEU94hN2qRw63djhEdsdsNqOwsBD19fXWDsWmKRQK+Pj4NHnKT3v4CU7URRKJBNOeCMHFq6VYnnIMZgFIJUBKYiwC+7tYOzwim1ZurER+mR5BqgA4+oRA6qC0dkhEdqewsBAuLi5wceFnTnvKy8tRWFiIAQMGdGh/llsQdZOsvFKYb90GaxbAv7OKYKgwWjcoIhuXWXIRLo4aaAtzofJnqQXR/aivr2eC3AEuLi6dGm1nkkzUTYYGe+D2/CJSCRAeoMPyPx3FB1+eQdHNausGR2SjMoovIMIjBMar2VAGRFo7HCIiC5ZbEHWTfjonpCTG4lzOdQwN9oCXToWlzw3Hzn9cwKL1BxE30g//ET8IPh4aa4dKZDPOF1/EE+6hECY+H5nIXphMJjz//PMAgMzMTISEhEChUGDx4sWIiYlp87gVK1Zg/fr1DyjKrmOSTNRNlI5yBPZ3aVKHHBXiiagQT2RcvoGdB7Lx0sbDeHy4L2Y+GYoBXs5WjJbI+u7UI/uzHpnoAag1NuDazWqcvTWYc783mMtkMmzfvh0AMHfuXKSkpECn0wFoTKBlMlmrx9lTggwwSSZ6IIYEuePtRQ8jK+8mPj9wAS9vOoxHon0xc1woAvqxjoz6pqySS3B21EBbeAVK1iMTdYvKmnoY6xparJdKJCitMGLZH49YbjB//7UnoNU4wtzKvHKODnJoVIoOXVOv12Px4sXw9fWFTqdDXFwctmzZgvr6egQFBWHdusYJgiZPnow9e/YgKSkJDg4OKCwshNFoRGpqKhwdHbv2wnsAk2SiByg8QIe3fvMQLl01YOeBbCxN/r94aGh/PPdkGIJ8Xa0dHtEDlVF8AYM9QmA8cQzaMc9YOxwiu2cymbHgnX+gurZlkhwZ7I7oQZ5NbjA/lXENZy6U4NzlGy32d1LKsWPNRMhkHbt9Ta/X469//StUKhWqq6stI82vvfYazpw5g+jo6Cb7DxkyBGvWrME777yDf/3rX4iLi+vkq+15TJKJrCDET4v/+l9jkFtYhs8PXMCyPx7BqMH98Ny4UAzyc7N2eEQPxPniC3iM9chE3UYmk2Lbf41vdyT5s++zLCPJY4b0w1NjAtocSe5oggw0Tu+sUqkAANnZ2fjTn/6E+vp6FBYW4qmnnmqRJEdERAAA+vfvj7Kyss68zAeGSTKRFQ30cUXS/FHIu1aOXQcvYHnKMQwP88KscWEID9RZOzyiHlNhrERemR5BStYjE3UnjUrRZpmEylHe5AZz726c9OruOuQPP/wQb7zxBiIiIrBs2TKIVpJwiURi+bq17baASTKRDQjo54LXfz0Ss8eH4YtDF/H7D44jKtgDz40LRWSwh7XDI+p2mbfrkX+5AhUf/Ub0QLR2g3lPmDBhApYvX46goCCbTYA7QiLsKPqwsDBkZ2dbOwyiHvfL9Sp8cegCDv90FREDdZg1LgxRIR5NfvMm+2evfVp3xP3fp3fhRnUppp84Bu8Zb8BpYPS9DyKiVuXm5mLgwIHWDsMuNP9etdefcTIRIhvU30ONpc8Nx0crnoSflzNWbf0Bv998HKeziu36t3Ki2zJKLiLUUQthaoDSl/XIRGR7mCQT2TAvnRMW/0c0tv7nkwge4Iq1fz2FxJRjSDt/jcky2a1KYxXyDXoMrKplPTIR2SzWJBPZAQ+tCoumRWFmfCh2H7mETdt/gq+nBrPGhWLMkP6QSlmGQfbjfMlFaByc4PbLFagC+HxkIrJNTJKJ7IibixILpkTi2dhB+D9HL+GPn52Gl5sTXno2CmqVAmdzbnRpFiWiB+F88QVEeIag7sQxuD001drhEBG1ip+iRHZI6+yI5ycPwfTYQTiYlgeTSeDV/31nFqWUxNgev3uZ6H5llFzEo24hECYT65GJyGaxJpnIjrmoHTA9dhDyrpU3mUUpLeMa9p+8grM511HfYLZqjER3u12PHFTNemQiezVnzhzk5+dblnNycjBv3rwW++3evRvbtm0DAKxYsaLF9lOnTmHNmjVtXufUqVO4dOmSZbm1c/QkJslEvcDQEE/cLkuWSoBhoZ7IuHwDqz46iYSV3+Gdj09h34lcXLtRZd1Aqc/LvH6J9chEdm7SpEnYu3evZXnv3r2YPHlyu8esX7++09dJS0trkiTfzzm6guUWRL1AP51Ti1mUEhN+hVdmDkNGzg38O7sI//PPy9jyVTp8PdUYEe6NEWFeiAx2h9KB3QA9OBnFFxDuEYy6k/9kPTLRA2auq0W9oQi1eRlQBgyBQut9X3/NmThxIhYsWICXXnoJAHDgwAEMGzYM3377LaqqqvDyyy8jPj6+yTGTJ0/Gnj17UFxcjMTERDg4OMDd3R0ajQYAsGnTJpw9e9Zy/OjRo/H1119DpVLhyy+/RGpqquUcOTk5eOutt2A2mxEUFIS3334baWlp+Oijj6BWq3H58mWsXLkSo0eP7tL3i5+ORL1AW7MoOSpkGBHuhRHhXsAzQNHNapzOLsbP2cXYtP0nNJjMiAxyb9wnzAt+3s6csIR61PniC3jELbixHnlAuLXDIep1TLVVEPXGlhskUpiqDNBvex0QZkAihe+CdyFTaxuXm++ucIRMqW71GjqdDjqdDjk5OTAajfD19cWKFSvg5OSEsrIyvPDCCy2S5Ns++ugj/PrXv8ZTTz2FzZs34+bNmwCAV155pcXx06ZNw6BBgzBhwoQm50hOTkZSUhIiIyOxatUqHDp0CM7OzqiurkZqairOnz+Pv/zlL0ySiajjvHVOmBgTiIkxgWgwmZF15SZOZxfjyOkCbPufDHhoVfjVrYQ5epAn1CqFtUOmXqSyrgp5Bj3mOgyA0ncQpApHa4dE1KsIswn5m1+EMFa32Kb0HwLVwKF3EmJhRvXFH1GTm47a/PMt9pc4OiHwtf+GRCpr9VqTJ0/G3r17UVtbi0mTJmH79u04cuQI5HI5CgsL24zxypUrWLhwIQAgKioKR44cAYAOHw8ABQUFiIxsnM5++PDhyM3NRVRUFMLDwyGRSNC/f38YDIZ2z9ERTJKJ+ii5TIrIYA9EBntg3tODUVpRi5+zS3A6qxgffHkGlTX1CA9ww4hwL/wqzBtBvq58HjN1SVbJJagdnKD75QqU/qxHJupuEqkM/q/8pd2R5NJjuywjyU6DRsF52Lg2R5LbSpABYNy4cZgzZw4aGhowf/58vPjii/jqq69QXl6Op59+us3jAgMDkZ6ejnHjxiE9PR0AYDAYsH///hbHKxQKmEymFufw9fXFuXPnEBkZiZ9//hmPPPJIY8zd/JdQJslEBABwc1YibqQf4kb6wWwWyNEbcDqrGP/OLMaO/VlwVjtgeGhj6cbwUC9onTkKSJ2TUXwREbfqkXUx06wdDlGvJFOqgTbKJKQOSvj+Jhm1eRlQBURCrvW67yfMaDQaBAQEQC6Xw8vLCz4+PkhISMDgwYPh7Ozc5nELFy5EYmIidu7ciX79+sHR0RGurq6tHh8TE4MNGzbgyJEjePfddy3nWL58OVauXAmgMemOi4vDjz/+eF+voz0SYUdz24aFhSE7O9vaYRD1OZU19ThzsXGU+XRWEa6X1SJkgCuGh3nh0WG+kAA4e5kTmXSWvfZp9xv3h2mfINDBGYP27UDg8r+z3IKom+Tm5mLgwIHWDsMuNP9etdef8ZOMiO5Jo1JgbJQPxkb5QAiB/KIKnM4qxqUCA8qr6rDywxOWiUzeXfoYhGisf3bVOPBGQEJtgxFFlSXwUOswyCsM6sdnMkEmIpvHJJmIOkUikSCgnwsC+jU+SePbf+Y0mcjk/10oRvqlGzhzsQRqpRw+nhr4empu/a+Gj6cGPh5qOCl5U2BfUVx5HW/8Yx2EENgl2YONsYnWDomI6J6YJBNRl9yeyOT2SPKYIf3xbFwoSkqrUVhSBX1JJQpLKpGZewMHf8xHSWk1hAB0LspbybMavpZEWg1vnRoKOec56k0yii/gdmWfEAKZN68g0DPYylEREbWPSTIRdUlrE5nIpBL0c1ejn7u68RnNd6mrN+GXG1UoLKmEvqTx/x/O/YLCkioYKo2QSiXw1jk1TaA9Gkei3V2VfMKGHRriFQqJRAIhBCQSCYZ48/nIRGT7mCQTUZe0NZFJWxwUsiblGnerrKlH4a2R59sJdMblGygsqUSN0QQHhQw+HmpLAh09yBMalQLnc29gaIgH+unUvGnQBnlpPPDu+Ddx7pcMDHYLhJfGw9ohERHdEz9NiMhmaFQKhPq7IdTfrcl6IQRKK4yW0g19SRUMFY2jzq+9f9RS6vH2iw/jq8OXIISA1tkRrhpHuN3631XjCK2zI7S3vmZJx4OjlDvCX+sL10tngepcKPuFWjskIqJ7YpJMRDZPIpFA56KEzkWJocF3RiG/PX65yU2Dl/VleHpsIK4UlsNQacR1Qw1yCspgqKyFoaIOFdV1lmPVKgW0GgdonZVw1ThAq7mVQN+VSN9OsJ2U8laf0lFrbMC1m9U4e6vUhI+/a525rhb1hiKYayvgFDwC5rra+342KxFZn8lkwvPPPw8AyMzMREhICBQKBRYvXoyYmJg2j8vMzERNTQ1GjBjxgCLtGvbmRGS3hgZ7NLlpcHioFwL6u2DMkP6t7t9gMqO8qg5llUaUVhhRVtn4z1BhhKHSiOz80sblyjoYKoxoMDXOQqWQSxtHou9Kqn081Ige5IU3/nzMcv2UxNgOl530JfWGIuhTlwPCjNJju+D7m2Q4egVYOyyiPqe2wYjiyuvIKL6AIV6h8NJ4QCnv/OMYZTIZtm/fDgCYO3cuUlJSoNPp7nlcZmYmSktLmSQDwK5du7B7924oFAqsW7cOfn5+lm3p6elYt67xkUAvvvgiYmNjezIUIuqFWrtpsD1ymdQyIn2vx+4LIVBd2wDDrSS6MXk2oqzCiNJKI+obzEi/VNJkJPtcznUmya2ozcu4M+2tMKM2L4NJMlEPqaqrhtFU12K9VCKBoaYCvz+wznIT7cbx/wmt0hnmVuaVc5Q5QO3Qfp96mxACK1asgF6vBwC8/fbb8PHxwdKlS1FRUQGJRIK1a9fi73//OyorK3HkyBFs2bKl3Zn5bEGPJckGgwFffPEFPvvsM5w/fx7JyclISUmxbF+/fj1SUlKg0WiQkJCAxx57DDJZ23OEExE119mbBjtDIpFArVJArVLA11PT6j5XfinHJ/syLSPJd5eC0B3KgCGARNqYKEukUAVEWjskol7JZDZh8Z43UVNf22LbYM9BGOod3uRxjP/WpyO9KAuZJRdb7K9SKPHx1GTIpPfOzXbu3ImIiAisX78e2dnZSElJwbJlyyCEwCeffAIAMJvNmDdvHkpLS7FgwYIuvtIHo8eS5PT0dIwePRpyuRxRUVHIzc21bDMajTCZTPD29gbQOO/2lStXEBzM52YSkf3o7Eh2X6XQesP3N8mozcuAKiAScq3XvQ8iok6TSWXYMnltuyPJuzL2WEaSR/pGIT54bJsjyR1JkAHg+vXrOHToEA4cOAAAUKlU8PPzw6OPPorExES4ubnhd7/7XddenBX0WJJcVlYGV1dXy7K4qwEMBkOTIXYXFxeUlZU1Of7Pf/4zNm/e3FPhERF1WU+OZPcmUgclHL0CWGJB9ACoHZygRuu/sCvlSrw7/k2cL76Iwd6h8FK731dNcnPu7u6YOXMmZs2aBQCoq6tDXV0dZs+ejYSEBGzZsgV79+6Fk5MTTCZTl6/3oPRYkuzi4oLs7GzLslR653FLrq6uqKiosCxXVFQ0SagBYMmSJViyZEmTdWFhYT0ULREREVHvdvtxjP5a32497+zZs/Hee+/hu+++gxAC8fHxePzxx/Hmm29CJpNBCIFNmzbBbDbj008/xblz57Bu3TpoNK2XstmKHkuSo6OjsWXLFphMJmRlZSEg4M4IglKphEwmQ3FxMTQaDfLy8ppsJyIiIiLbd/spF2vXrm2xbceOHS3W7dy5s8dj6i49liRrtVpMnToVCQkJkMvlWLt2LXbv3o0BAwZg9OjRSEpKwtKlSyGEwMsvvwy5nE+jIyIiIiLb0KOZ6ezZszF79mzL8t2jxdHR0Xb12wQRERER9R2cl5WIiIiIqBkmyURERER2TKFQoLy83Nph2Lzy8nIoFIoO789CYCIiIiI75uPjg8LCQty4ccPaodg0hUIBHx+fDu/PJJmIiIjIjkmlUgwYMMDaYfQ6LLcgIiIiImqGSTIRERERUTN2V27BWfeIiKyPfTER9XqiD5g2bZpVz9XRY+61X3vb29rW2vrW1oWGht4jup7TXe3Dtul+vaFt7rWPPbePvent76fuWM+f967t11t/1tk27a/vqbZhuQURERERUTOyVatWrbJ2EA9CZGSkVc/V0WPutV9729va1tr65us2b96MJUuWdCDCntFd7cO26X69oW3utY89t4+96e3vp66ut/b7qTe0T2/9WWfbtL2+p9pGIoQQ3X5WsjthYWHIzs62dhjUCraNbWP7UHfi+8l2sW1sV0+1TZ8ZSaZ7GzNmjLVDoDawbWwb24e6E99PtottY7t6om04kkxERERE1Axv3CMiIiIiaoZJMhERERFRM0ySiYiIiIiasbsZ96jnXbp0CampqXjiiScwYcIEa4dDzZw4cQLnzp1DXl4eXn/9dWi1WmuHRLdkZmbihx9+QH5+PpYsWQKdTmftkMiOsS+2beyLbVd39cUcSe4j6uvrMWvWLIwcORL79++3rN+1axdmzZqFuXPn4urVqwCAkJAQTJs2zVqh9kmdaZ+HH34Yv/3tb+Ht7Y2KigprhdxndKZtIiIioNVqUVRUBJlMZq2QyYaxL7Zt7IttlzX6Yj4Cro+QSqV4/PHHIZfL4e7ujpCQEBgMBmzYsAE7duxAaGgoPvjgA0ycOBEAoNfrUVtbi5CQECtH3jd0tn12794NT09PjBw50sqR936dbZuIiAio1WrU1dXB29vbytGTrWFfbNvYF9sua/TFHEnuIyQSCby8vJqsS09Px+jRoyGXyxEVFYXc3FwAQFFREb7//nscPXoU+fn51gi3z+lM+3zxxRf4/vvvUVJSAr1eb41w+5TOtM2hQ4ewdetWHD58mAkytYp9sW1jX2y7rNEXsya5DysrK4Orq6tl+fYjs729vbFy5UprhUW3tNU+M2bMwIwZM6wVFqHttomPj0d8fLy1wiI7xb7YtrEvtl093RdzJLkPc3FxQXl5uWVZKuXbwZawfWwX24a6E99Pto3tY7t6um3Y0n1YdHQ0fvzxR5hMJmRkZCAgIMDaIdFd2D62i21D3YnvJ9vG9rFdPd02LLfoQ1599VWcO3cOTk5OSE9PxxtvvIGpU6ciISEBcrkca9eutXaIfRrbx3axbag78f1k29g+tutBt41E3C7gICIiIiIiACy3ICIiIiJqgUkyEREREVEzTJKJiIiIiJphkkxERERE1AyTZCIiIiKiZpgkExERERE1wySZrGLs2LFNlpOSknDs2DEAwMKFC1FXV9fimGXLluHUqVMt1icnJ2P37t1djqmurg5r167Fk08+ienTp2P+/PnIyspCQUEBZs6c2eXz34+4uDgYjUarXJuIejf2wx3Dfrjv4mQiZHO2bt1qlesmJyejtrYW+/fvh1wuR15eHvR6PTQajVXiISKyFvbDREySyQbFxcVh3759cHR0xPvvv499+/bB19cXd897s2PHDvztb3+Du7s7PDw8EBQUBABIT0/Hhg0bUFNTAz8/P2zcuBEqlQpjx47FxIkTcfLkSfj7+2Pz5s2QyWSW81VXV+Obb77B4cOHIZc3/lgEBAQgICAABQUFqK+vR2JiIjIyMjBmzBisXr0aAPCHP/wBGRkZMBqNmDNnDhISEgCgzevFxcVh6tSpOHjwIJydnfHhhx9Co9HgypUrWL16NcrKyuDm5oaNGzfCw8PDEl9RURFeffVV1NTUwGw24/3330dwcHCPtwUR9U3sh9kPEwBBZAWDBw8WU6ZMsfwbNWqUOHr0qBBCiNjYWFFbWyvOnDkjZsyYIerq6sS1a9fEsGHDxA8//CCuXbsmxo8fLyoqKkRFRYWIjY0VX331lTAajSIhIUEYDAYhhBBbt24VqampQgghQkNDRVpamhBCiEWLFonjx483iSczM1M888wzrcZ69epVERkZKXJzc0VDQ4OYMmWKyM3NFUIIUVpaKoQQoq6uTkyfPl3cuHGj3evFxsaKr7/+WgghxJo1a8Tnn38uhBDihRdeEAUFBUIIIb777juxevXqJt+Lbdu2iZSUFMu1ampq7vt7T0QkBPthIdgPU/s4kkxWodVq8c0331iWk5KSWuxz+vRpjBs3DgqFAt7e3hg5ciQA4OzZs4iJibH8+e2xxx4DAOTm5iI7Oxvz5s0DANTX1yMmJgYA4OLiglGjRgEAIiIioNfrW1xPIpG0GW9wcDACAwMBAGFhYdDr9QgMDMSePXvw5ZdfwmQy4dq1a8jLy4NOp2v3evHx8Zb1V69eRWVlJU6fPo3FixcDAMxmM/z8/Jpcf+jQoUhKSoJUKsWECRMQEhLSZqxERB3Bfpj9MLWPSTLZtLs7zHt9LYTA0KFD8fHHH7c4j4ODg+VrqVQKk8nUZLu/vz8KCwtRXV0NJyenDh1/9epVfP755/jss8+g0WiwaNEiy40u7V3v9rbb64UQ8Pb2bvJh1dyoUaPw6aef4vDhw1iyZAlWrlxp+eAhIupJ7IcbsR/ue/h0C7JZI0aMwMGDB1FfX4+ioiL89NNPABp/mz9x4gQqKytRWVlpuRs7KCgIBQUFyM7OBtBY35aXl9ehazk5OWHKlCnYuHGjpSPNz8/HyZMn2zymqqoKarUaarUaBQUFSEtLu6/X6ezsDBcXFxw/fhxA48hLTk5Ok330ej08PT0xZ84cTJo0yfIaiYh6EvvhO9gP9z0cSSabFRUVhYceegiTJ0/GgAEDMGzYMACAt7c35s+fj2effRbu7u6IjIwE0DgykJycjLfeegvV1dUQQuD1119HQEBAh663fPlybNq0CePHj4eTkxPc3d2xYsWKNvcPDw+Hn58fJkyYAH9/f4wYMeK+X2tycjJWrVqFTZs2wWQyYcGCBU1uCElLS0NqaioUCgW0Wi3ee++9+74WEVFHsR9mP9yXSYS461ZVIiIiIiJiuQURERERUXNMkomIiIiImmGSTERERETUDJNkIiIiIqJmmCQTERERETXDJJmIiIiIqBkmyUREREREzTBJJiIiIiJq5v8DGVAzG9VnrscAAAAASUVORK5CYII=\n"
          },
          "metadata": {}
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   Hidden Channels    Loss  Train  Validation    Test\n",
              "0                2  0.9125  0.735      0.5325  0.5431\n",
              "1                4  0.5606  1.000      0.7425  0.7541\n",
              "2                8  0.4704  1.000      0.7575  0.7826\n",
              "3               16  0.3232  1.000      0.7700  0.7992\n",
              "4               32  0.1995  1.000      0.7850  0.8103\n",
              "5               64  0.1104  1.000      0.7925  0.8106\n",
              "6              128  0.0605  1.000      0.7925  0.8127\n",
              "7              256  0.0320  1.000      0.7975  0.8130\n",
              "8              512  0.0169  1.000      0.7825  0.7980\n",
              "9             1024  0.0080  1.000      0.7800  0.7912"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-5f346ed1-1ed0-4d0c-8e9b-1ec2642c3c9e\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Hidden Channels</th>\n",
              "      <th>Loss</th>\n",
              "      <th>Train</th>\n",
              "      <th>Validation</th>\n",
              "      <th>Test</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>2</td>\n",
              "      <td>0.9125</td>\n",
              "      <td>0.735</td>\n",
              "      <td>0.5325</td>\n",
              "      <td>0.5431</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>4</td>\n",
              "      <td>0.5606</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.7425</td>\n",
              "      <td>0.7541</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>8</td>\n",
              "      <td>0.4704</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.7575</td>\n",
              "      <td>0.7826</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>16</td>\n",
              "      <td>0.3232</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.7700</td>\n",
              "      <td>0.7992</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>32</td>\n",
              "      <td>0.1995</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.7850</td>\n",
              "      <td>0.8103</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>64</td>\n",
              "      <td>0.1104</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.7925</td>\n",
              "      <td>0.8106</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>128</td>\n",
              "      <td>0.0605</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.7925</td>\n",
              "      <td>0.8127</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>256</td>\n",
              "      <td>0.0320</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.7975</td>\n",
              "      <td>0.8130</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>512</td>\n",
              "      <td>0.0169</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.7825</td>\n",
              "      <td>0.7980</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>1024</td>\n",
              "      <td>0.0080</td>\n",
              "      <td>1.000</td>\n",
              "      <td>0.7800</td>\n",
              "      <td>0.7912</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-5f346ed1-1ed0-4d0c-8e9b-1ec2642c3c9e')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-5f346ed1-1ed0-4d0c-8e9b-1ec2642c3c9e button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-5f346ed1-1ed0-4d0c-8e9b-1ec2642c3c9e');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 59
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def gcn_num_plot(gcn_nums, max_epoch=100, save=\"gcn_nums.png\"):\n",
        "    data = []\n",
        "    for gnum in gcn_nums:\n",
        "        model = gcn(hidden_channels=64, num_gcn=gnum, num_lin=1)\n",
        "        optimizer = torch.optim.Adam(model.parameters(), lr=0.005)\n",
        "        values = []\n",
        "        for epo in range(max_epoch):\n",
        "            loss = train(model, optimizer)\n",
        "            accs = test(model)\n",
        "            values.append([epo+1, loss, *accs])\n",
        "        data.append([gnum, *values[-1][1:]])\n",
        "    df = pd.DataFrame(data, columns=[\"GCN Layers\", \"Loss\", \"Train\", \"Validation\", \"Test\"])\n",
        "    fig, ax = plt.subplots(1, 2, figsize=(10, 5))\n",
        "    sns.lineplot(data=df, x=\"GCN Layers\", y=\"Loss\", marker=\"o\", ax=ax[0])\n",
        "    sns.lineplot(data=df, x=\"GCN Layers\", y=\"Train\", label=\"Train\", marker=\"o\", ax=ax[1])\n",
        "    sns.lineplot(data=df, x=\"GCN Layers\", y=\"Validation\", label=\"Validation\", marker=\"o\", ax=ax[1])\n",
        "    sns.lineplot(data=df, x=\"GCN Layers\", y=\"Test\", label=\"Test\", marker=\"o\", ax=ax[1])\n",
        "    ax[1].set(ylabel=\"Accuracy\")\n",
        "    fig.savefig(save)\n",
        "    plt.show()\n",
        "    return df.round(4).head(None)\n",
        "\n",
        "\n",
        "gcn_num_plot(range(1, 13))  "
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 778
        },
        "id": "fYkUGuKdRdI0",
        "outputId": "d71a1c5f-71c1-4409-8bbc-7190118249bc"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x360 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "    GCN Layers    Loss   Train  Validation    Test\n",
              "0            1  0.1109  1.0000      0.7625  0.7777\n",
              "1            2  0.1104  1.0000      0.7925  0.8106\n",
              "2            3  0.1138  1.0000      0.7975  0.8198\n",
              "3            4  0.1209  1.0000      0.7825  0.8093\n",
              "4            5  0.1185  1.0000      0.7775  0.8241\n",
              "5            6  0.1258  1.0000      0.8075  0.8385\n",
              "6            7  0.1239  1.0000      0.7950  0.8121\n",
              "7            8  0.1220  1.0000      0.7975  0.8155\n",
              "8            9  0.1518  0.9950      0.7825  0.8333\n",
              "9           10  0.5627  0.8225      0.6950  0.7602\n",
              "10          11  0.6678  0.6725      0.5800  0.5800\n",
              "11          12  0.5251  0.7425      0.6900  0.6727"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-e2f8b4a3-500a-4ccc-9e42-7b269a5b0283\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>GCN Layers</th>\n",
              "      <th>Loss</th>\n",
              "      <th>Train</th>\n",
              "      <th>Validation</th>\n",
              "      <th>Test</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1</td>\n",
              "      <td>0.1109</td>\n",
              "      <td>1.0000</td>\n",
              "      <td>0.7625</td>\n",
              "      <td>0.7777</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2</td>\n",
              "      <td>0.1104</td>\n",
              "      <td>1.0000</td>\n",
              "      <td>0.7925</td>\n",
              "      <td>0.8106</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>3</td>\n",
              "      <td>0.1138</td>\n",
              "      <td>1.0000</td>\n",
              "      <td>0.7975</td>\n",
              "      <td>0.8198</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4</td>\n",
              "      <td>0.1209</td>\n",
              "      <td>1.0000</td>\n",
              "      <td>0.7825</td>\n",
              "      <td>0.8093</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>5</td>\n",
              "      <td>0.1185</td>\n",
              "      <td>1.0000</td>\n",
              "      <td>0.7775</td>\n",
              "      <td>0.8241</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>6</td>\n",
              "      <td>0.1258</td>\n",
              "      <td>1.0000</td>\n",
              "      <td>0.8075</td>\n",
              "      <td>0.8385</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>7</td>\n",
              "      <td>0.1239</td>\n",
              "      <td>1.0000</td>\n",
              "      <td>0.7950</td>\n",
              "      <td>0.8121</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>8</td>\n",
              "      <td>0.1220</td>\n",
              "      <td>1.0000</td>\n",
              "      <td>0.7975</td>\n",
              "      <td>0.8155</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>9</td>\n",
              "      <td>0.1518</td>\n",
              "      <td>0.9950</td>\n",
              "      <td>0.7825</td>\n",
              "      <td>0.8333</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>10</td>\n",
              "      <td>0.5627</td>\n",
              "      <td>0.8225</td>\n",
              "      <td>0.6950</td>\n",
              "      <td>0.7602</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>11</td>\n",
              "      <td>0.6678</td>\n",
              "      <td>0.6725</td>\n",
              "      <td>0.5800</td>\n",
              "      <td>0.5800</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>12</td>\n",
              "      <td>0.5251</td>\n",
              "      <td>0.7425</td>\n",
              "      <td>0.6900</td>\n",
              "      <td>0.6727</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-e2f8b4a3-500a-4ccc-9e42-7b269a5b0283')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-e2f8b4a3-500a-4ccc-9e42-7b269a5b0283 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-e2f8b4a3-500a-4ccc-9e42-7b269a5b0283');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 64
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def gcn_lin_plot(lin_nums, max_epoch=100, save=\"gcn_lin.png\"):\n",
        "    data = []\n",
        "    for lnum in lin_nums:\n",
        "        model = gcn(hidden_channels=64, num_gcn=3, num_lin=lnum)\n",
        "        optimizer = torch.optim.Adam(model.parameters(), lr=0.005)\n",
        "        values = []\n",
        "        for epo in range(max_epoch):\n",
        "            loss = train(model, optimizer)\n",
        "            accs = test(model)\n",
        "            values.append([epo+1, loss, *accs])\n",
        "        data.append([lnum, *values[-1][1:]])\n",
        "    df = pd.DataFrame(data, columns=[\"FC Layers\", \"Loss\", \"Train\", \"Validation\", \"Test\"])\n",
        "    fig, ax = plt.subplots(1, 2, figsize=(10, 5))\n",
        "    sns.lineplot(data=df, x=\"FC Layers\", y=\"Loss\", marker=\"o\", ax=ax[0])\n",
        "    sns.lineplot(data=df, x=\"FC Layers\", y=\"Train\", label=\"Train\", marker=\"o\", ax=ax[1])\n",
        "    sns.lineplot(data=df, x=\"FC Layers\", y=\"Validation\", label=\"Validation\", marker=\"o\", ax=ax[1])\n",
        "    sns.lineplot(data=df, x=\"FC Layers\", y=\"Test\", label=\"Test\", marker=\"o\", ax=ax[1])\n",
        "    ax[1].set(ylabel=\"Accuracy\")\n",
        "    fig.savefig(save)\n",
        "    plt.show()\n",
        "    return df.round(4).head(None)\n",
        "\n",
        "\n",
        "gcn_lin_plot(range(1, 13))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 778
        },
        "id": "T66L1H0xUxJV",
        "outputId": "cc64224f-6d92-4bcc-b79a-5c648f46dee7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x360 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "    FC Layers    Loss   Train  Validation    Test\n",
              "0           1  0.1138  1.0000      0.7975  0.8204\n",
              "1           2  0.0359  1.0000      0.7875  0.8124\n",
              "2           3  0.0364  1.0000      0.7800  0.7983\n",
              "3           4  0.0068  1.0000      0.7825  0.7792\n",
              "4           5  0.2545  0.9925      0.6900  0.7172\n",
              "5           6  0.1060  1.0000      0.7125  0.7157\n",
              "6           7  0.0806  1.0000      0.6700  0.7025\n",
              "7           8  0.3639  0.8275      0.6275  0.6488\n",
              "8           9  0.0627  1.0000      0.6550  0.6985\n",
              "9          10  0.2674  1.0000      0.7100  0.7246\n",
              "10         11  0.2268  1.0000      0.6800  0.7221\n",
              "11         12  0.1787  0.8300      0.6300  0.6795"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-ff6a1f81-c0b9-4e09-9ac2-63d92f784f75\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>FC Layers</th>\n",
              "      <th>Loss</th>\n",
              "      <th>Train</th>\n",
              "      <th>Validation</th>\n",
              "      <th>Test</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1</td>\n",
              "      <td>0.1138</td>\n",
              "      <td>1.0000</td>\n",
              "      <td>0.7975</td>\n",
              "      <td>0.8204</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2</td>\n",
              "      <td>0.0359</td>\n",
              "      <td>1.0000</td>\n",
              "      <td>0.7875</td>\n",
              "      <td>0.8124</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>3</td>\n",
              "      <td>0.0364</td>\n",
              "      <td>1.0000</td>\n",
              "      <td>0.7800</td>\n",
              "      <td>0.7983</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4</td>\n",
              "      <td>0.0068</td>\n",
              "      <td>1.0000</td>\n",
              "      <td>0.7825</td>\n",
              "      <td>0.7792</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>5</td>\n",
              "      <td>0.2545</td>\n",
              "      <td>0.9925</td>\n",
              "      <td>0.6900</td>\n",
              "      <td>0.7172</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>6</td>\n",
              "      <td>0.1060</td>\n",
              "      <td>1.0000</td>\n",
              "      <td>0.7125</td>\n",
              "      <td>0.7157</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>7</td>\n",
              "      <td>0.0806</td>\n",
              "      <td>1.0000</td>\n",
              "      <td>0.6700</td>\n",
              "      <td>0.7025</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>8</td>\n",
              "      <td>0.3639</td>\n",
              "      <td>0.8275</td>\n",
              "      <td>0.6275</td>\n",
              "      <td>0.6488</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>9</td>\n",
              "      <td>0.0627</td>\n",
              "      <td>1.0000</td>\n",
              "      <td>0.6550</td>\n",
              "      <td>0.6985</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>10</td>\n",
              "      <td>0.2674</td>\n",
              "      <td>1.0000</td>\n",
              "      <td>0.7100</td>\n",
              "      <td>0.7246</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>11</td>\n",
              "      <td>0.2268</td>\n",
              "      <td>1.0000</td>\n",
              "      <td>0.6800</td>\n",
              "      <td>0.7221</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>12</td>\n",
              "      <td>0.1787</td>\n",
              "      <td>0.8300</td>\n",
              "      <td>0.6300</td>\n",
              "      <td>0.6795</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-ff6a1f81-c0b9-4e09-9ac2-63d92f784f75')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-ff6a1f81-c0b9-4e09-9ac2-63d92f784f75 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-ff6a1f81-c0b9-4e09-9ac2-63d92f784f75');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 65
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model = gcn(hidden_channels=64, num_gcn=3, num_lin=1)\n",
        "optimizer = torch.optim.Adam(model.parameters(), lr=0.005)\n",
        "print(\"GCN:\")\n",
        "eval(model, optimizer, max_epoch=100)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Qo-2gqvetq1i",
        "outputId": "836aff4a-d39f-4c06-ae31-246708f46b4c"
      },
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "GCN:\n",
            "Accuracy: 0.8204\n",
            "Computation Time: 14.5502s\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Part b, GAT\n",
        "\n",
        "def gat(hidden_channels=64, heads=1):\n",
        "    torch.manual_seed(1234567)\n",
        "    layers = [\n",
        "        (GATConv((-1, -1), hidden_channels, heads=heads, add_self_loops=False), 'x, edge_index -> x'),\n",
        "        ReLU(inplace=True),\n",
        "        (GATConv((-1, -1), hidden_channels, heads=heads, add_self_loops=False), 'x, edge_index -> x'),\n",
        "        ReLU(inplace=True),\n",
        "        (GATConv((-1, -1), num_classes, heads=heads, add_self_loops=False), 'x, edge_index -> x')\n",
        "    ]\n",
        "    model = Sequential('x, edge_index', layers)\n",
        "    model = to_hetero(model, data.metadata()).to(device)\n",
        "    return model\n",
        "\n",
        "def gat_sub(hidden_channels=64, heads=2):\n",
        "    torch.manual_seed(1234567)\n",
        "    layers = [\n",
        "        (GATConv((-1, -1), hidden_channels, heads=heads, add_self_loops=False, return_attention_weights=True), 'x, edge_index -> x'),\n",
        "        ReLU(inplace=True),\n",
        "        (GATConv((-1, -1), hidden_channels, heads=heads, add_self_loops=False, return_attention_weights=True), 'x, edge_index -> x'),\n",
        "        ReLU(inplace=True),\n",
        "        (GATConv((-1, -1), num_classes, heads=heads, add_self_loops=False, return_attention_weights=True), 'x, edge_index -> x'),\n",
        "        (Linear(heads * num_classes, num_classes), 'x -> x'),\n",
        "    ]\n",
        "    model = Sequential('x, edge_index', layers)\n",
        "    model = to_hetero(model, data.metadata()).to(device)\n",
        "    return model"
      ],
      "metadata": {
        "id": "Evm1o0JUbg7l"
      },
      "execution_count": 55,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def gat_head_plot(head_nums, max_epoch=100, save=\"gat_head.png\"):\n",
        "    data = []\n",
        "    for i in head_nums:\n",
        "        model = gat(heads=i)\n",
        "        optimizer = torch.optim.Adam(model.parameters(), lr=0.005)\n",
        "        values = []\n",
        "        for epo in range(max_epoch):\n",
        "            loss = train(model, optimizer)\n",
        "            accs = test(model)\n",
        "            values.append([epo+1, loss, *accs])\n",
        "        data.append([i, *values[-1][1:]])\n",
        "    df = pd.DataFrame(data, columns=[\"Heads\", \"Loss\", \"Train\", \"Validation\", \"Test\"])\n",
        "    fig, ax = plt.subplots(1, 2, figsize=(10, 5))\n",
        "    sns.lineplot(data=df, x=\"Heads\", y=\"Loss\", marker=\"o\", ax=ax[0])\n",
        "    sns.lineplot(data=df, x=\"Heads\", y=\"Train\", label=\"Train\", marker=\"o\", ax=ax[1])\n",
        "    sns.lineplot(data=df, x=\"Heads\", y=\"Validation\", label=\"Validation\", marker=\"o\", ax=ax[1])\n",
        "    sns.lineplot(data=df, x=\"Heads\", y=\"Test\", label=\"Test\", marker=\"o\", ax=ax[1])\n",
        "    ax[1].set(ylabel=\"Accuracy\")\n",
        "    fig.savefig(save)\n",
        "    plt.show()\n",
        "    return df.round(4).head(None)\n",
        "\n",
        "\n",
        "gat_head_plot(range(1, 11))"
      ],
      "metadata": {
        "id": "LaOoK538c8WN",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 716
        },
        "outputId": "1d25590c-dc20-4697-9ec7-7198d20a0ba3"
      },
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x360 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   Heads  Loss  Train  Validation    Test\n",
              "0      1   0.0    1.0      0.8850  0.9054\n",
              "1      2   0.0    1.0      0.9275  0.8907\n",
              "2      3   0.0    1.0      0.8925  0.8987\n",
              "3      4   0.0    1.0      0.9125  0.8968\n",
              "4      5   0.0    1.0      0.8675  0.8667\n",
              "5      6   0.0    1.0      0.8800  0.8827\n",
              "6      7   0.0    1.0      0.8475  0.8652\n",
              "7      8   0.0    1.0      0.8225  0.8652\n",
              "8      9   0.0    1.0      0.8700  0.8870\n",
              "9     10   0.0    1.0      0.8575  0.8652"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-15e456a0-91b1-4c80-b23e-82e23e5522d4\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Heads</th>\n",
              "      <th>Loss</th>\n",
              "      <th>Train</th>\n",
              "      <th>Validation</th>\n",
              "      <th>Test</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.8850</td>\n",
              "      <td>0.9054</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.9275</td>\n",
              "      <td>0.8907</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>3</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.8925</td>\n",
              "      <td>0.8987</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.9125</td>\n",
              "      <td>0.8968</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>5</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.8675</td>\n",
              "      <td>0.8667</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>6</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.8800</td>\n",
              "      <td>0.8827</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>7</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.8475</td>\n",
              "      <td>0.8652</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>8</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.8225</td>\n",
              "      <td>0.8652</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>9</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.8700</td>\n",
              "      <td>0.8870</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>10</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.8575</td>\n",
              "      <td>0.8652</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-15e456a0-91b1-4c80-b23e-82e23e5522d4')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-15e456a0-91b1-4c80-b23e-82e23e5522d4 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-15e456a0-91b1-4c80-b23e-82e23e5522d4');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 41
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model = gat(hidden_channels=64, heads=2)\n",
        "optimizer = torch.optim.Adam(model.parameters(), lr=0.005)\n",
        "print(\"GAT:\")\n",
        "eval(model, optimizer, max_epoch=100)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pjpT_bc44yLO",
        "outputId": "65a3d198-7e90-4ec0-f5af-a6511d5bdb95"
      },
      "execution_count": 57,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "GAT:\n",
            "Accuracy: 0.8907\n",
            "Computation Time: 12.7384s\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model = gat_sub(hidden_channels=64, heads=2)\n",
        "optimizer = torch.optim.Adam(model.parameters(), lr=0.001)\n",
        "print(\"GAT using learnable coefficient:\")\n",
        "eval(model, optimizer, max_epoch=100)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YfXFU_RCpV5k",
        "outputId": "9a5d33f6-ea23-4f2a-f76d-34d5c8a3f311"
      },
      "execution_count": 56,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "GAT using learnable coefficient:\n",
            "Accuracy: 0.9116\n",
            "Computation Time: 12.6774s\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "class GATv2(GATv2Conv):\n",
        "    def __init__(self, *args, **kwargs):\n",
        "        self.operator = kwargs.pop(\"operator\")\n",
        "        super().__init__(*args, **kwargs)\n",
        "\n",
        "    def message(self, x_j, x_i, edge_attr, index, ptr, size_i):\n",
        "        if self.operator == \"Hadamard\":\n",
        "            x = x_i * x_j\n",
        "        elif self.operator == \"Min\":\n",
        "            x = torch.minimum(x_i, x_j)\n",
        "        elif self.operator == \"Max\":\n",
        "            x = torch.maximum(x_i, x_j)\n",
        "        else:\n",
        "            x = x_i + x_j\n",
        "        x = F.leaky_relu(x, self.negative_slope)\n",
        "        alpha = (x * self.att).sum(dim=-1)\n",
        "        alpha = softmax(alpha, index, ptr, size_i)\n",
        "        self._alpha = alpha\n",
        "        alpha = F.dropout(alpha, p=self.dropout, training=self.training)\n",
        "        return x_j * alpha.unsqueeze(-1)\n",
        "\n",
        "\n",
        "def gatv2(hidden_channels=64, heads=2, operator=\"Hadamard\"):\n",
        "    torch.manual_seed(1234567)\n",
        "    layers = [\n",
        "        (GATv2((-1, -1), hidden_channels, heads=heads, add_self_loops=False, operator=operator), 'x, edge_index -> x'),\n",
        "        ReLU(inplace=True),\n",
        "        (GATv2((-1, -1), hidden_channels, heads=heads, add_self_loops=False, operator=operator), 'x, edge_index -> x'),\n",
        "        ReLU(inplace=True),        \n",
        "        (GATv2((-1, -1), num_classes, heads=heads, add_self_loops=False, operator=operator), 'x, edge_index -> x')\n",
        "    ]\n",
        "    model = Sequential('x, edge_index', layers)\n",
        "    model = to_hetero(model, data.metadata()).to(device)\n",
        "    return model\n"
      ],
      "metadata": {
        "id": "8cb1CNYKoJVP"
      },
      "execution_count": 58,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model = gatv2(operator=\"Hadamard\")\n",
        "optimizer = torch.optim.Adam(model.parameters(), lr=0.001)\n",
        "print(\"GATv2 using Hadamard:\")\n",
        "eval(model, optimizer, max_epoch=100)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NB4eNswgrOff",
        "outputId": "1e14ed6f-26ef-4e11-fbb3-8b72517e1cdb"
      },
      "execution_count": 59,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "GATv2 using Hadamard:\n",
            "Accuracy: 0.9143\n",
            "Computation Time: 15.5159s\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model = gatv2(operator=\"Min\")\n",
        "optimizer = torch.optim.Adam(model.parameters(), lr=0.001)\n",
        "print(\"GATv2 using Min:\")\n",
        "eval(model, optimizer, max_epoch=100)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8LwWYswjroMy",
        "outputId": "693ad4a8-0596-499b-fc37-b56585540886"
      },
      "execution_count": 60,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "GATv2 using Min:\n",
            "Accuracy: 0.9183\n",
            "Computation Time: 17.0286s\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model = gatv2(operator=\"Max\")\n",
        "optimizer = torch.optim.Adam(model.parameters(), lr=0.001)\n",
        "print(\"GATv2 using Max:\")\n",
        "eval(model, optimizer, max_epoch=100)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TM6KbSznr5F7",
        "outputId": "c828c46c-da45-44c2-e606-c53a17902948"
      },
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "GATv2 using Max:\n",
            "Accuracy: 0.9245\n",
            "Computation Time: 17.0818s\n"
          ]
        }
      ]
    }
  ]
}